The

Future of
the Electric
Grid
AN INTERDISCIPLINARY MIT STUDY

Other Reports in This Series

The Future of Nuclear Power (2003)

The Future of Geothermal Energy (2006)

The Future of Coal (2007)

Update to the Future of Nuclear Power (2009)

The Future of Natural Gas (2011)

The Future of the Nuclear Fuel Cycle (2011)

This study can also be viewed online at
http://web.mit.edu/mitei/research/studies/the-electric-grid-2011.shtml

Copyright © 2011 Massachusetts Institute of Technology. All rights reserved.

ISBN 978-0-9828008-6-7

ii MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

Study Participants
STUDY CO-CHAIRS

JOHN G. KASSAKIAN WILLIAM W. HOGAN
Professor of Electrical Engineering and Raymond Plank Professor of Global Energy Policy
Computer Science, MIT Research Director, Harvard Electricity Policy Group,
Former Director, MIT Laboratory for Mossavar-Rahmani Center for Business and
Electromagnetic and Electronic Systems Government, John F. Kennedy School
of Government, Harvard University
RICHARD SCHMALENSEE
Howard W. Johnson Professor of HENRY D. JACOBY
Economics and Management, MIT William F. Pounds Professor of Management Emeritus,
Former John C Head III Dean, MIT Sloan MIT
School of Management
JAMES L. KIRTLEY
STUDY MANAGEMENT Professor of Electrical Engineering, MIT

GARY DESGROSEILLIERS HARVEY G. MICHAELS
Executive Director Research Scientist, Department of Urban Studies
and Planning, MIT
TIMOTHY D. HEIDEL
Research Director IGNACIO PÉREZ-ARRIAGA
Postdoctoral Associate, MIT Energy Initiative Professor of Electrical Engineering
Comillas University, Spain
STUDY GROUP Visiting Professor, Engineering Systems Division, MIT

KHURRAM AFRIDI DAVID J. PERREAULT
Associate Professor and Werner-von-Siemens Professor of Electrical Engineering and Computer
Chair for Power Electronics, Lahore University Science, MIT
of Management Sciences, School of Science
and Engineering, Pakistan NANCY L. ROSE
Visiting Associate Professor of Electrical Charles P. Kindleberger Professor of Applied Economics,
Engineering and Computer Science, MIT MIT

AMRO M. FARID GERALD L. WILSON
Assistant Professor of Engineering Systems and Vannevar Bush Professor of Electrical Engineering
Management, Masdar Institute, UAE Emeritus, MIT
Research Affiliate, Technology and Development Former Dean, MIT School of Engineering
Program, MIT

JERROLD M. GROCHOW
Research Affiliate, MIT Energy Initiative
Former Vice President for Information Services
and Technology, MIT

MIT Study on the Future of the Electric Grid iii

STUDENT RESEARCH ASSISTANTS

NABI ABUDALDAH VIVEK A. SAKHRANI
Wageningen University, The Netherlands Engineering Systems Division, MIT
Visiting Student, MIT
JIANKANG WANG
MINJIE CHEN Electrical Engineering and Computer Science,
Electrical Engineering and Computer Science, MIT
MIT
ANDREW WHITAKER
PEARL E. DONOHOO Engineering Systems Division, MIT
Engineering Systems Division, MIT
XIANG LING YAP
SAMANTHA J. GUNTER Harvard University
Electrical Engineering and Computer Science, Engineering Systems Division, MIT
MIT
RICHARD Y. ZHANG
P. JORDAN KWOK Electrical Engineering and Computer Science,
Engineering Systems Division, MIT MIT

iv MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

Advisory Committee Members
J. BENNETT JOHNSTON — CHAIR STEVEN NAUMANN
United States Senate (Retired) Exelon Corporation

GEORGE ARNOLD PEDRO PIZARRO AND DAVID MEAD
National Institute of Standards and Technology Southern California Edison

LISA M. BARTON CLAES RYTOFT
American Electric Power ABB Group

TROY BATTERBERRY AND MIGUEL ÁNGEL SÁNCHEZ FORNIÉ
ELIZABETH GROSSMAN Iberdrola S.A.
Microsoft Corporation
BASEM SARANDAH
WILLIAM W. BERRY Nexant, Inc.
Dominion Resources (Retired)
J. CHARLES SMITH
CLARK GELLINGS Utility Wind Integration Group
Electric Power Research Institute
SUSAN TIERNEY
ROBERT GILLIGAN AND LARRY SOLLECITO Analysis Group
General Electric Company
GORDON VAN WELIE
LAURA IPSEN AND PAUL DE MARTINI ISO New England
Cisco Systems
STEPHEN G. WHITLEY
PAUL JOSKOW New York ISO
Alfred P. Sloan Foundation

ELIZABETH ANNE MOLER
Former Chair of the Federal Energy Regulatory
Commission

MIT Study on the Future of the Electric Grid v

vi MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .

. cial.A. economy and the quality of American life. support. thorough discussions of key topics to serve as able system of systems will continue to serve us references. we linking electric generation units reliably and mainly consider the implications of a set of efficiently to millions of residential. The while we do not make assumptions regarding U. opportunities that the grid will face. electricity has made predecessors focused on implications of vital contributions to the growth of the U. the challenges and opportunities it the challenges and opportunities it is likely to will face. Chapter 1 provides an overview of the status objective portrait of the U. American Electric Power. of the contents of each section in Chapters 2–9 focused research and demonstration. Instead. ongoing trends and existing policies. it will face serious challenges in the next mendations. and the are provided in each chapter’s introduction. we include appen- dices on the grid’s history and technology. and Southern California considers many technologies and multiple Edison. our efforts were sponsors gave us access to staff members who focused on integrating and evaluating existing provided frequent and detailed information knowledge rather than performing original about technical and policy issues. to support our findings and recom- well.000 organizations. This report aims to provide a comprehensive. Also.S. It also highlights facilitate selective reading.S. our study of the grid necessarily and Technology. Microsoft focused on particular technologies and energy Corporation. To face over the next two decades. collection and sharing of important data can and recommendations are collected and briefly facilitate meeting the challenges and seizing the discussed in each chapter’s final section. government as they guide the grid’s continuing many of which are in turn regulated by both evolution. have shed light on a range of complex and Bechtel Foundation. familiar with the industry. General environment. In addition. Its predecessors study: ABB Group. While this remark. for those less more appropriate regulatory policies.S. and industrial users of electricity through more than six million miles of lines and We anticipate this report will be of value to a associated equipment that are designed and wide range of decision makers in industry and managed by more than 3. Iberdrola S. We are very research and analysis. While the previous studies have Electric Company. electric grid and of the grid. In addition to providing financial overlapping physical and regulatory systems. and our major recommendations. National Institute of Standards supply. national policies limiting carbon emissions. Larry Birenbaum. commer. future carbon policy initiatives. electric grid is a remarkable achievement. We have attempted to provide federal and state agencies. MIT Study on the Future of the Electric Grid vii . detailed descriptions a number of areas in which policy changes.Foreword and Acknowledgments For well over a century. The MIT Future of the Electric Grid Study This study is the sixth in the MIT Energy gratefully acknowledges the sponsors of this Initiative’s “Future of” series. Exelon Corporation. this study’s grateful for this cooperation. and to meet the needs of what we two decades that will demand the intelligent expect will be a diverse audience in terms of use of new technologies and the adoption of interest and expertise. Cisco important issues involving energy and the Systems. many of our corporate and government Because of this breadth.

In addition. from the sponsors. Moniz. helped identify members of the Advisory Committee. the Advisory Committee. We would especially like to acknowledge the efficient conduct of The final report represents the opinions of the Advisory Committee meetings under the able study group. selected the Co-Chairs of the study. content of the report. Matt Dinsmore. Natalie Liang. Rebecca Marshall-Howarth. and helped engage the study sponsors. are not responsible for. Emily Fisher. Any errors in the final docu- available experts from their own organizations ment are the responsibility of the study group to answer our questions and contribute to the and not the editor or anyone else. Andrew Bochman. The Director of MITei. we would like to thank Sarah Aldy for a significant amount of their time to participate editing this document and maintaining her in multiple meetings. The Advisory Committee members dedicated Finally. research on this project also benefited from individual conversations with Lauren Azar. and Justin Daniels. and other members of their respective organi- zations. Chair. Professor Ernest J. and make cated iterations. Joseph Eto. Richard O’Neill. This study was initiated and performed within the MIT Energy Initiative (MITei). and countless others who we interacted with at conferences and industry events. viii MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . MITei staff provided administrative and financial management assistance to this project. The advisory committee and sponsors Johnston. which is solely responsible for its and experienced direction of Senator J. a number of individuals in the Brattle Group and National Grid. Bennett content. Paul Centolella. Arun Phadke. Joseph Hezir. we would like to acknowledge the important contributions of Melanie Kenderdine. read and comment on patience with our team through many compli- several early drafts of the report. the findings and recommendations In addition to all of the valuable contributions contained within this report. Patricia Connell. and do not necessarily endorse.

S.5 Conclusions and Recommendations 53 Chapter 3: Integration of Variable Energy Resources 55 3.1 Opportunities in Distribution System Operation 132 6.4 Interconnecting Variable Energy Resources 71 3.3 Ensuring Adequate System Flexibility 69 3.1 The Transmission Network and System Operations 36 2.4 Concluding Remarks 31 Chapter 2: Enhancing the Transmission Network and System Operations 32 2.5 Conclusions and Recommendations 77 Chapter 4: Transmission Expansion 79 4.1 Characteristics of Variable Energy Resources 56 3.4 Improving System Operations 49 2. and Major Recommendations 2 1. 83 4.2 Transmission Planning 88 4.1 Today’s Electric Grid 11 1.Table of Contents xi Abstract 1 Chapter 1: Challenges.5 Conclusions and Recommendations 109 Chapter 5: The Impact of Distributed Generation and Electric Vehicles 109 5.3 Pilot Programs and Deployment Challenges 139 6.2 Electric Vehicles 122 5.1 Distributed Generation 116 5.3 Increasing Transmission Capacity 46 2.3 Conclusions and Recommendations 127 Chapter 6: Enhancing the Distribution System 128 6.3 Transmission Cost Allocation 97 4.4 Conclusions and Recommendation MIT Study on the Future of the Electric Grid ix .1 Transmission Development in the U.2 Advanced Metering for the Distribution System 137 6.4 Siting New Transmission Capacity 101 4.2 Preventing Blackouts 39 2.2 Variable Energy Resources and the Cost of Reserves 64 3. Opportunities.3 Major Recommendations 28 1.2 Challenges and Opportunities 21 1.

4 Conclusions and Recommendations Appendices 235 Appendix A: A Brief History of the U.2 Cybersecurity of the Electric Grid 219 9.3 How Should Policy Respond? 193 8.2 Demand Response Programs Today 156 7.3 Interconnection and Competition 243 Appendix B: Electric Power System Basics 243 B.3 Structure of the Electric Power System 254 B.4 Operation of the Electric Power System 258 B.3 Predicting the Benefits of Increased Demand Engagement 161 7.1 Why Engage Demand? 148 7.1 Beyond Municipal Boundaries 236 A.1 Introduction 243 B.6 Conclusions and Recommendations 175 Chapter 8: Utility Regulation 176 8. and Information Privacy 199 9.5 Expanding Demand Engagement: Findings 167 7.2 Growing Challenges for Regulatory Policy 185 8.4 Reducing Residential Energy Consumption 164 7.S.4 Conclusions and Recommendations 197 Chapter 9: Data Communications. Grid 235 A.1 Grid Data Communications 208 9.1 Regulatory Objectives and Processes 179 8. Cybersecurity.2 Fundamentals of Electric Power 247 B.6 Power System Planning 261 Glossary 267 Acronyms and Abbreviations x MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .3 Information Privacy and Security 227 9.2 A Federal Role Emerges 238 A. 143 Chapter 7: Engaging Electricity Demand 145 7.5 Wholesale Electricity Markets 259 B.

A failure to realize these opportunities or meet these challenges Increased penetration of electric vehicles and could result in degraded reliability. renewable generation in response to policy and increase flexibility in controlling power flows initiatives at both state and federal levels. penetration of renewable distributed genera- The grid will face a number of serious challenges tion will pose challenges for the design and over the next two decades. increase the public policy goals.S. and make possible renewable generators to be located far from the engagement of demand as a resource. ratio of peak to average demand and thus further reduce capacity utilization and raise This report.Abstract The U. regional. could help to a comprehensive. These technologies can enhance efficiency and One of the most important emerging challenges reliability. operation of distribution systems. for the design and operation of distribution focused research. communica- guide the grid’s necessary evolution. New regulatory approaches may be sharing can contribute to meeting the chal. and may gies also present valuable opportunities for raise costs for many consumers. a linked system of public and private enterprises and siting regimes will need to be changed to operating within a web of government institu. if measures are not taken. described above. facilitate this expansion. often via human network connecting thousands of unusually long transmission lines. Changes in retail pricing policies. enabled Initiative’s Future of series. integrating and evaluating existing knowledge rather than performing original research. If properly deployed and of this capacity will rely on either solar or wind accompanied by appropriate policies. they can power and will accordingly produce output that deal effectively with some of the challenges is variable over time and imperfectly predictable. control. aims to provide by new metering technology. significantly other ongoing changes in electricity demand increased costs. increase capacity utilization. Increased penetration electric grid and the identification and analysis of distributed generation will pose challenges of areas in which intelligent policy changes. while new technolo. Much on transmission lines.S. and a failure to achieve several will. In addition. provide greater visibility of the the best resource locations will require many instantaneous state of the grid. electric grid is a vast physical and expansion of the transmission system. meeting these challenges. existing load centers and will thus necessitate MIT Study on the Future of the Electric Grid xi . objective portrait of the U. and data development and systems. enable facing the grid is the need to incorporate more more rapid response to remediate contingencies. increased tions: federal. They can facilitate the integra- making it harder for system operators to match tion of large volumes of renewable and distrib- generation and load at every instant. and municipal. cost-allocation procedures. Current electricity generators to millions of consumers— planning processes. and power electronics. required to encourage the adoption of innova- lenges the grid is facing. We Opportunities for improving the functioning hope it will be of value to decision makers in and reliability of the grid arise from techno- industry and in all levels of government as they logical developments in sensing. the fifth in the MIT Energy rates. state. It reflects a focus on tive network technologies. tions. mitigate this problem. Utilizing uted generation.

have taken important actions in recent years to guide the evolution of the U. and privacy. and thus they raise incentives related to distributed generation important issues of standardization. recover fixed network costs through customer charges that do not vary with the volume of Decision makers in government and industry electricity consumption. cyber. electric power s 4O MAKE EFFECTIVE USE OF NEW TECHNOLOGIES. All these new technologies involve increased s 4O IMPROVE UTILITIES AND THEIR CUSTOMERS data communication. and energy conservation. utilities should security.S.

and a number tools for bulk power system operation. and models of consumer summarized as follows: response to real-time pricing. and technical methods for wide-area transmission planning. procedures for response to and recovery from Our main recommendations can be briefly cyberattacks. regulatory. the Federal Energy Regulatory ingly complex and dynamic environment. Yet the diversity of owner. and recovery across the entire electric power sector. of institutional. the electric power industry should fund nities noted above. several key areas. xii MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . s 4O FACILITATE THE INTEGRATION OF REMOTE s 4O IMPROVE DECISION MAKING IN AN INCREAS- renewables. system to address the challenges and opportu. comprehensive results from “smart grid” demonstration projects. power system. including information on the bulk that cross state lines. increased research and development in ship and regulatory structures within the U. Commission should be granted enhanced more detailed data should be compiled and authority to site major transmission facilities shared. a single federal agency performance. including computational grid complicates policy-making. should be given responsibility for cybersecu- rity preparedness. utilities with advanced metering technology should begin a transition to pricing regimes in which customers pay rates that reflect the time-varying costs of supplying power. response.S. and s 4O COPE MORE EFFECTIVELY WITH INCREASING standardized metrics of utility cost and cybersecurity threats. including both bulk power and distribution systems. impediments remain that require action. s 4O IMPROVE THE GRIDS EFlCIENCY AND LOWER rates.

owned. electric as used in this report refers not only to the power grid serves more than 143 million physical transmission and distribution systems residential. meeting emerging workforce needs. handling the charging of electric vehicles.) Section 1. commercial. research and development. regulatory. Section 1. distributed electric generators. cybersecurity and privacy. forthcoming.3 then presents the major recommendations that flow from these concerns. “Creating a Smarter U.S. Joskow. public policies and a variety of and control the physical elements. regulatory and governance structures that ment of the 20th century” by the National shape the system’s evolution.S. and the need for improved data development and sharing of information.S. this is likely to be a period of slow growth in or load. Opportunities. (Readers may consult Appendix A for a brief history of the U.S.S. Two less tangible elements are also the U. electric grid over the next two decades— energy—up from 14% in 1949—underscoring a period long enough to permit significant the great and growing national importance of change but short enough to make it unlikely the grid’s efficiency and reliability. electric generation This study considers the evolution of the consumed 41% of the nation’s primary U. and the technological and economic changes will alter i An excellent brief overview of many of the issues considered in this report is provided by P. system.2 then outlines challenges and opportunities that the grid will confront over the next two decades: incorporating variable energy sources like wind and solar. high-voltage transmission.S. Chapter 1: Challenges. organized by area of industry and government policy response: the transmission system.1 the U. government-owned. Section 1. Hailed as the “supreme engineering achieve. and Major Recommendations This chapter gives an overview of the study and introduces our major findings and recommendations. lower.Chapter 1: Challenges. the distribution system. L. summarizes the history of the U.S. adjusting distribution systems to accommodate small-scale. Electricity Grid. and industrial that link generators to ultimate loads but also customers 2 through more than 6 million miles the associated operational. grid and Appendix B for additional information on the technical operation of electric power systems. making the best use of new technologies to ensure reliability and efficiency under changing conditions. Appendix A briefly by more than 3. and cooperative enterprises.4 that unforeseen technologies will This study considers the evolution The electric power system is composed of have significant of the U.3 In 2009. which serve as initial conditions for our analysis. Opportunities.” Journal of Economic Perspectives. generation. grid. and of transmission and distribution lines owned governance structures.000 highly diverse investor.i Even though voltage distribution. The term “grid” Academy of Engineering.1 provides a brief description of the grid’s current structure and performance.4 concludes with brief remarks about the level of urgency attached to these issues. and responding to issues presented by the vast increase of data communications within the grid. Along the way some of the study’s key findings are summarized. demand for electricity by historical important: the operational systems that protect standards. and Major Recommendations 1 . and energy consumption. Section 1. electric grid over the four interacting physical elements: energy impacts on the next two decades.

necessary knowledge of the workings of electric power systems. limits on transmis- the technologies employed in the grid do not sion line capacities. ii As we discuss in later chapters. The federal government primarily has added new policies on top of old ones. does not have a comprehensive national electricity policy. operations. Power systems networks subject to incentive-based regulation.S. sector. Several features of electric power systems are In natural gas. and fundamental to their structure and operation: telecommunications—industries with a history of deep and ongoing government involvement— s )T IS GENERALLY NOT ECONOMICALLY FEASIBLE TO federal policy was substantially reformed after store electricity in bulk. Even though state boundaries do not affect the flow of s %LECTRICITY mOWS THROUGH MANY TRANSMISSION electricity and thus have no natural role in the paths from generators to customers. individual flows cannot be controlled with state regulators retain considerable authority. both the demand for and supply of electricity s 4HE NEED TO MAINTAIN INSTANTANEOUS BALANCE in challenging ways. and regulatory s !N ELECTRIC GRID IS A NATURAL MONOPOLY regimes differ substantially among states. to consult Appendix B. research and development is performed. trucking.ii Technologies exist that can meet these chal- lenges effectively. it would be prohibitively expensive to have multiple overlapping grids in any region. but some parts of later chapters may be The U. this last requirement has implications for many policies.1 TODAY’S ELECTRIC GRID important data are compiled and shared. including the role and design of competitive electricity markets. If regulatory policies and of supply and demand. precision. of the U. The widespread 1970 to reflect market realities. unlike s )T FOLLOWS THAT ELECTRIC POWER SYSTEMS MUST the European Union and many other nations vary the supply of electricity to meet minute. deployment of new storage technologies and/ despite dramatic changes in the electric power or high penetrations of electric vehicles may sector. and design or operation of the electric power sector. does not have a comprehensive national hard to understand without more background. that have adopted comprehensive new struc- to-minute changes in demand and in the tures based on competitive wholesale and retail output of variable energy sources such as electricity markets and centrally managed wind and solar generators. and other features require change. but only if a number of This chapter assumes essentially no additional regulatory policies are changed. railroads. and 1. electric power industry.S. The U. it is likely to be difficult to maintain central coordination of short-term system acceptable reliability and electric rates. airlines. federal policies established in the 1930s someday change this. we first discuss the organization capacity for safety. and regulatory regimes differ Readers who would like more background or encounter material they find difficult are urged substantially among states. In contrast. electricity policy.5 2 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . must be built with enough capacity to meet expected peak demand with some excess In this section.S. but these developments and even earlier still play a central role in that are unlikely before 2030.

gov/sites/prod/files/oeprod/DocumentsandMedia/NERC_Interconnection_1A. They are linked by only a few performance. consists of three cooperative entities are important in some independently synchronized grids: the Eastern regions but not at all in others.The result is substantial regional differences. organized the most part we take today’s policy regime as wholesale markets do not exist in large parts of given. At the highest level. as Section 1. and 8%. respectively.pdf. we believe the nation. elec- not conducive to efficiency.S. account for marks.1. Figure 1.1 Interconnections of the North American Electric Grid Source: U. These well relative to available international bench. and Major Recommendations 3 . Interconnection. shown in Figure 1. the electric power system Subsidies of various sorts for public and of the continental U. adequate support for fundamental policy reform may be For a variety of reasons. Because the grid is tricity sales. three grids. despite working within a policy regime 73%. http://energy. 19%.6 currently functioning well. Structure Organized wholesale markets for power are central in some areas and nonexistent in others. but. Department of Energy. The industry has done reasonably low-capacity direct current (dc) lines. Chapter 1: Challenges.3 makes clear. despite national policy supporting that some policy changes are necessary to competitive wholesale markets with open. however. and the Electric Reliability Council of Texas We then turn to a discussion of the industry’s (ERCOT). the Western Interconnection. discussion of which unlikely to emerge in the near term. prevent a deterioration of the grid’s perfor- mance in light of emerging challenges. Thus for are beyond the scope of this study.S. Opportunities.S. of U.

the U. particularly the Pacific lines. this and demand for power in real time in specified policy has major implications for the transmis. electric grid currently retail customers.org. own parts of the transmission or bulk power system. http://www. non-discriminatory access to transmission which are responsible for balancing the supply systems (see Figure 1. 14%.S.400 miles of 765 kV alternating current (ac) while in the West. municipally owned. These markets now cover consists of approximately 170.isorto. which do not reflect differences in levels of supply or are operated by Independent System Operators demand or in system complexity. all rights reserved. it strong influence of history. (ISOs) or Regional Transmission Organizations (RTOs) that do not own generators or serve Physically. Nonetheless. In the Southeast.. the traditional distribution lines. while Arkansas and Arizona each have eight and Figure 1. 2. Figure 1. population and meet high-voltage (above 200 kilovolts or kV) electric about two-thirds of U.S.7 provides an overview of the operation of these and almost 6 million miles of lower-voltage markets.3. federal. and federal enterprises own balancing authorities. Copyright © ISO/RTO Council. A glance at this map makes plain the sion grid.2). shown in Figure 1.8 These include approximately vertically integrated utility model is dominant. Box 1. and the U. and Texas are each tightly integrated common elements. the highest voltage lines in operation in Northwest.10 Investor-owned utilities own about Within these broad areas are 107 so-called 66% of the system.1 transmission lines and associated equipment. areas. demand.2 shows the geographic scope of Florida has eleven. New York.9 Several hundred entities currently in the industry. The rest is divided among other publicly 4 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .000 miles of 500 kV cooperative enterprises play an important role dc lines.000 miles of two-thirds of the U. New has led to organized markets with important England. and more than 3. These differences plainly organized wholesale electricity markets. and have one balancing authority each.S.S.2 Regions with Organized Electricity Markets Source: ISO/RTO Council. Where it has been implemented.

are costly to shut down or bring back on price higher than the contract the seller pays line. Federal Energy use. load and the state of the generator fleet— owned entities (7%). in part because these large The U. the ISO or RTO functions electricity in the U. For a reliability concerns. but contracted sales incorporating such parameters as the time are settled by a side transaction between the required to start the generator.1 WHOLESALE ELECTRICITY MARKETS though such a swing is not a usual occurrence. is around $100/MWh. While zero bids from renewable price ($/MWh). and 27% of electricity and others (3%).S. out-of- seller and buyer that accounts for the real-time economic-order dispatch due to congestion or price the parties paid and received. these categories of customers independent transmission companies (4%). real-time price below the contract price the Some base-load generators.S. cooperatives (6%). and will offer their energy at a price of zero the difference to the buyer. respectively. In some cases of very light load these generators In addition to energy.11 Chapter 1: Challenges.000 industrial customers.) as both the operator of the system and the Some generators that use renewable energy— financial exchange for wholesale energy sales. where a buyer and called the “clearing price”—the offer of the last a seller agree on a fixed price over a set period.6 million commercial and incur lower delivery costs. which is designed to energy can swing wildly during a day—from ensure that there is enough “iron in the ground” near zero to near $1. Opportunities. and security constraints. All the generators that are The majority of energy is traded through dispatched receive the same compensation long-term bilateral contracts. customers. customers pay about 8% less per kilowatt tion over wholesale electricity sales and hour (kWh) for electricity than residential transmission rates. market as described above. The U. matches load. the electricity market has may offer their energy at a negative price to a number of other “products. to ensure that they are always dispatched.” the treatment of guarantee they remain on line since the cost of which varies from region to region. 17. grid serves about 125 million resi. such as nuclear buyer pays the seller the difference. and Major Recommendations 5 . while industrial customers pay about 40% less. such as reserves of energy price. customers.S. the average retail price of electricity markets. the ISO/RTO in ascending order called the they do not necessarily result in savings to “bid stack” and the generators are dispatched society because the true cost of the renewable (told to generate) in this order until generation energy is the clearing price plus the subsidy.000/MWh depending on to meet future needs. generator dispatched. account for 37%. 36%. wind and solar mainly—have essentially zero Selling wholesale electric energy begins with variable costs. customers can take power at higher voltages dential customers. The actual process is Energy is still bought and sold in the real-time more complicated than this simple explanation. On average. Consequently the cost of wholesale various types and capacity. and 775. In those areas of the country with wholesale (For comparison. and for a plants. and the subsidies they receive for a bidding process whereby generators offer generation permit them to bid a negative price an amount of energy (MWh) for sale during and still receive positive compensation for their specific periods of the next day at a specific energy. These offers are arranged by generators reduce the market clearing price. In aggregate. Among stopping and starting outweighs the negative these are ancillary services. BOX 1. commercial Regulatory Commission (FERC) has jurisdic.

In some areas of the country. about 7.com/docs/oc/rs/BubbleMap_2011-04-12.200 organizations rates are regulated by state public utility commissions (PUCs). Their retail At the distribution level. but they account for 66% of electricity sales.3 Balancing Authorities in the North American Electric Grid.5% of provide electricity to retail customers. Finally. 2011 Source: North American Electric Reliability Corporation. municipally tions provide electricity to retail customers. about 3. http://www. Another 818 are coopera. which in aggregate account for 10.ORTHEAST 0OWER #OORDINATING #OUNCIL 2&#  2ELIABILITY&IRST #ORPORATION 3%2#  3%2# 2ELIABILITY #ORPORATION 300  3OUTHWEST 0OWER 0OOL 42%  4EXAS 2EGIONAL %NTITY WECC = Western Electricity Coordinating Council.200 are publicly owned—six by the substantially from preferred access to low-cost federal government and the rest by states and power from federal projects. Only 242 distribu- municipalities—but they account for only 16% tion entities are investor owned. services. tives.OTE &2##  &LORIDA 2ELIABILITY #OORDINATING #OUNCIL -2/  -IDWEST 2ELIABILITY /RGANIZATION .5% of kWh sales.0##  . about 3. 6 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .12 owned and cooperative utilities benefit Nearly 2.jpg .nerc.Figure 1. retail sales are accounted for by retail power marketers that do not provide distribution of electricity sold.200 organiza. At the distribution level. particularly the Pacific Northwest.

investor-owned and other dimensions of performance. rates of growth. in technical publications. Currently. These enterprises mainly in the late 1920s to less than 7% today.14 unattainable at any cost. the development and deployment of more efficient transformers and FINDING other equipment. are difficult active retail choice programs for residential because of differing geography. but only in Texas do more and definitions of performance measures.In some states (and in much of Europe).S. and transmission at higher As a result of the layering of historical policy voltages.S. Moreover.S. and of other components. or government-owned) supplier However. because there District of Columbia more than 60% of large are diminishing returns to investing to increase commercial and industrial customers have efficiency and reliability. That fraction has quence of the structural reforms described in fallen significantly over time in the U. distribution systems. from a competitive supplier. Chapter 1: Challenges. and Major Recommendations 7 . with the distribu. decisions and the lack of a comprehensive. iii Losses are measured as the difference between energy generated and energy delivered to customers and thus in practice include losses due to theft. but it is significant in some other nations.S. assessing the performance of a system of distribution services. electric grid is not a or no retail competition of this sort in 35 states. grid is often referred to as “anti- tion utility remaining as the owner of the quated” or “broken” in the popular press and. The U.S. electricity generation. Publicly owned fraction of energy generated that is lost due to systems organized at the state or municipal heating of transmission and distribution lines level accounted for another 8%. shared vision of system structure or function. As Appendix A. today.16 cooperative.13 In those same for instance.iii This operated in regions with organized wholesale reflects investments in transmission and markets. As a conse. electricity customers. Theft is not considered to be important in the U. An important measure of the performance including the Tennessee Valley Authority. it is possible not just to underinvest but also to overinvest in these At the generation level. on average will have newer equip- jurisdictions. Opportunities. distribution network and the sole (regulated. utilities accounted for 42% of U.S. commercial and industrial ment. and in at least nine states and the driven by vendor R&D. losses in transmission and by independent power producers that do not distribution decreased from more than 16% serve retail customers.S. and perfection is switched to competitive suppliers. Comparisons over time may reveal customers can choose among multiple nothing more than the advance of technology suppliers. there is little as complex as the U. International comparisons and Fifteen states and the District of Columbia have even comparisons within the U. in 2007. occasionally. the remaining 42% was provided Figure 1. the U. Performance some customers can purchase electricity from competing retail suppliers.4 shows.15 Cooperatives and federal systems. electric power system today operates under a fragmented and often inconsistent policy regime. than 15% of those customers purchase power Systems that have grown more rapidly recently. each of a transmission and distribution system is the accounted for about 4%. simple task.

Figure 1.5 Population Density (ppl/km2) 2.5 Transmission and Distribution Losses for Selected Countries.5 2010 Cents Nominal Cents 0 100 200 300 400 500 0. Company and producer use data were not reported during this time period. Transmission and Distribution Losses. 2010 Cents Nominal Cents 50 50 50 50 Figure 1. 2008 9 Transmission and Distribution Losses 8 Canada United Kingdom 7 (% of Total Generation) Australia Transmission and Distribution Losses Italy France 6(% of Total Generation) United States China 5 Spain Germany Japan South Korea 4 3 4. Data for all years from 1951 to the present are from the U.org/indicator.S. net imports.0 0.0Bank Development Indicators. producer use.4 U.0 1 3.5 1. Losses for the years 1943 to 1951 were calculated using Edison Electric Institute data on generation.S. company use. and sales to customers. Source: World 1. 1926 to 2009 18 Transmission and Distribution Losses 16 14 (% of Total Generation) 12 10 8 6 4 2 0 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 Source: Data for all years prior to 1943 were reported by the Edison Electric Institute.worldbank.0 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 8 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . Energy Information Administration Annual Energy Reviews. http://data. so the average of these quantities from 1941 and 1942 was used.5 2 Israel 4.5 0 3.0 0 100 200 300 400 500 2.

S. expect. grid is in require utilities to report data on the impact line with that of other developed countries. or optimal. and the United Kingdom. or necessarily accurate. of all outages on consumers. since 1984. However. for reliability of the bulk power system. instance. Opportunities.17 (which correlates with the degree of urbaniza- This is on par with most European countries. national circumstances will clearly affect performance. so counts of comprehensive nor consistent. particularly its changes Data on outages are neither comprehensive nor consistent. Such comparisons cannot reveal whether where customers generally experience from less U. Increases in transmission voltage have been reported to the U. cannot reliably be used to assess changes in the cally in less than a few seconds. but only 35 U. less visible techno. In particular.S. urban areas averaging between 30 seconds and FINDING five minutes per year versus between nine hours Data are not available to quantitatively and almost four days in rural areas. too high. outages cannot be usefully where losses due to theft are said to be unusu.S. as one would countries. and several As a result of these and other advances. Lightning reliability of the bulk power system over time. losses are compa. Chapter 1: Challenges. losses tend to decline somewhat with increasing and between different Data on outages are neither population density.S.21 arrestors allowed the effects of lightning strikes to be contained automatically. Energy (DOE) since the 1970s and to the North logical advances have contributed to the grid’s American Electric Reliability Corporation reliability.S. reliability is too low. European nations.19 and accurately assess the reliability of the U. and Major Recommendations 9 . Most outages occur within over time. breakers and relaying allowed transmission complete.6 provides a comparison of minutes of outage per year in the U.S.S. and differences in density. At the same time. electric grid. these data are not consistent. Protective relaying enabled the (NERC).S. the data for Italy. Department of and many other significant. is not out of customers in the U. Figure 1. with power unavailable in U.S. but data are may often be even more important than nonetheless incomplete. and high-speed reclosing circuit However. the U.5 and 2 power interruptions per year account for differences in population density and between 2 and 8 hours without power. but it cannot indicate sons across space or over time. which has outage per customer year is not much affected a relatively old grid. data on Another important dimension of performance major disturbances and unusual occurrences is reliability. the whether U. than 1 interruption per year to almost 3. losses are higher or lower than treatment of very short interruptions varies would be optimal.18 There given the benefits of reducing outages and the is great variation between reliability in urban costs of doing so.20 It is accordingly systems that also have substantial amounts of impossible to make comprehensive compari- older equipment. what data are available distribution systems.Figure 1.5 indicates that U. can expect to experience line with other industrialized nations when we between 1. Minutes of ally high. however. tion). This figure also suggests that between states in the U. At the bulk power level. indicate that other factors by these definitional differences.S. all else equal.S. and rural areas. compared. and reporting rable to those of other wealthy countries with standards and practices differ. which has responsibility for the detection and isolation of system faults. and they lines to be reenergized after a fault automati. Here. states indicate the reliability of the U.S.

ATIONAL . Figure 1.H. Electric Power System: An Assessment of Publicly Available Information Reported to State Public Utility Commissions (Berkleley.ABORATORY.S.6 Average Duration of Interruptions for Selected Countries. Eto and K. CA: Lawrence Berkeley .H. 2006 160 United States 140 Duration of Interruptions 120 Spain (Minutes/Year) 100 United Kingdom 80 France 60 Italy 40 20 Germany 0 0 50 100 150 200 250 300 Population Density (ppl/km2) Source: United States Reliability Data: J. Lacommare. Tracking the Reliability of the U.

  %UROPEAN 2ELIABILITY $ATA #OUNCIL OF %UROPEAN %NERGY 2EGULATORS.

4th Benchmarking Report on Quality of Electricity Supply 2008 "RUSSELS.

"ELGIUM.

at an annual rate of 2.   0OPULATION $ENSITY 7ORLD "ANK $EVELOPMENT )NDICATORS 2010 Cents Nominal Cents 0 100 200 300 400 500 A final dimension of performance involves the beginning in the 1990s.5 on its suppliers Productivity improvement in the electric power for innovation. 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 The U. reluc- (% of Total for innovation. in part. output per labor hour rose 1. The lack of a long-term through the Electric Power Research Institute decline in later years suggests that the difference 2. collaborative Bureau of Labor Statistics shows that over the projects and toward shorter-term proprietary 1987 to 2008 period. but comparisons to other 2. as compared to 2. instead relying primarily 4. In recent industries remain relatively favorable: the shifted away1. utilities have sometimes tance among utilities to incur—and regulators to approve—R&D expenditures as federal and The U.7 shows a steady decline in collabora- devoted a very small fraction of its revenues to tive R&D spending through EPRI. as and have participated 3. investor-owned utilities.S. Generation) U.23 0. electric utility industry has historically Figure 1.22 Moreover. devoted a very small fraction of its revenues to R&D.5with vendors on R&D activities decline in real retail prices until the 1970s.5from longer-term. spending on average use of new technology to increase productivity.4% in power generation which account 0.5 for almost all nonfederal utility and supply.0 in collaborative research shown in Appendix A. This is reflected in a collaborated3. less than 1% of their revenues on R&D.0 10 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . The decrease R&D. instead relying Transmission primarily onLosses and Distribution its suppliers in utility R&D funding reflects. particularly during the 1990s.0 industry has historically been rapid relative to most other industries.0 years. electric utility industry has historically state policies pursued more industry competi- tion. 4.S. however. a nonprofit consortium founded in may have decreased.5 (EPRI). utilities have 1973.1% in the overall R&D spending. reduced their R&D budgets private nonfarm business sector.0 efforts.S.

since the mid-1980s. Figure 1. comparisons do not seem possible.0 and important challenges posed by other incentives for investment in renewable deployment of some of those technologies. electric grid is not broken Even though ment has supported the deployment of renew- today. and quantitative international supported in detail in later chapters. the federal govern- 3. initiative to reduce carbon dioxide emissions. the grid will face between now and 2030. could able energy sources—particularly geothermal. this differ. ence or its effects. Pollitt.0 Chapter 1: Challenges. 50 50 50 50 Data on broader measures of productivity do findings presented below are developed and not seem to exist. emerging 3. tax section briefly 2. the though support has been inconsistent over 1.0 challenges. however. But without better data there is world. federal and state subsidies and 4.S. Million (2003 Prices) 500 400 300 200 100 0 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Energy Conversion Energy Delivery and Utilization Environment Strategic Development Activities Source: T.7 Collaborative Research in the U. Opportunities. through the Electric Power Research Institute 700 600 $U. substantially degrade the system’s reliability biomass. we do not assume a carbon-constrained counterparts.5 0.0 Beginning in the late 1970s.1. and Major Recommendations 11 . Anecdotal 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 evidence from some vendors suggest that U.24 All states now provide tax credits or effectively. and solar—through accelerated 2.S.5 technologies that can be used to deal with them time.S.2 CHALLENGES AND OPPORTUNITIES of low-carbon generating technologies. Even in the absence of a broad federal no way to verify. This and efficiency depreciation and. if not met. have been less willing to deploy new Unlike prior studies in the MIT “Future of ” technologies in recent years than their foreign series. let alone quantify.25 and 29 states and the District of 0. “Liberalisation and R&D in Network Industries: the Case of the Electricity Industry. Jamasb and M. wind.5 regulations are tilting the playing field in favor 1.5 over the next few decades.5 the U.0 introduces the main challenges credits for either production or investment. particularly those that are investor owned.” Research Policy 37 (2008): 995–1008. Renewable Generation utilities.S. 4. The energy.

As Figure 1.27 A number of states have set very within the system will become increasingly ambitious requirements for renewables expan. cloud-free and sparsely populated desert gies are labeled “variable energy resources.” or Southwest. markets or for operating in a flexible manner.26 There is Devising and deploying mechanisms to enormous regional variation: in California. for over half of the non-hydro increase. Similarly.9 relates to concentrated solar power generation. locations for solar power are in the nearly For this reason.30 planning tools are inadequate for wide-area In particular as Chapter 3 notes. existing transmission variability if reliability is to be maintained. and wind and solar will account the Dakotas to the Canadian border. flexibility to power systems with organized iv Figure 1. at some cost. as shown in Figure 1.9. unlike centers. south. in which direct sunlight heats a working fluid (oil. important as the penetrations of wind and sion: in California renewables other than small solar generation increase. the renewables other than hydropower will account most attractive wind resources are in the “wind for 57% of the increase in generation between belt” that stretches north from Texas through 2010 and 2030. renewables other than large hydro facilities accounted for 13. the costs of offshore wind installations most other generating technologies. demand minus VER fossil-fueled or nuclear generating plants built generation (that is. as must be modified. The system and its operation compromise system stability. but.S. which is then used to generate steam to power a turbine. the prime ably over time and is imperfectly predictable. At low levels of penetration. generation in 2010. 12 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . Columbia have renewable portfolio standards. as Chapter 3 discusses and Exploiting these variable energy resources will several European systems have experienced.8 indicates.2% of U.iv VERs. The U. many of the most promising sites for not increased. few incentives planning. the net load that must be relatively close to load centers were driving met by other generators) becomes noticeably system expansion. First. Renewables other than hydropower accounted FINDING for 4.7% of electricity generation and for operating flexibly supply. but current federal policies are wind and solar generators are located far from simply continued beyond their sunset dates. the output are generally considerably greater than onshore of wind and solar generators varies consider. major load centers. to handle this Chapter 4 explains. and current cost-allocation methods exist today for investments that add generation need improvement.29 also has significant offshore wind potentials on both the East and West Coasts.S. offshore resources are closer to major load tial problems for the electric grid. for provide incentives for investment in flexible instance. facilities in good locations. hydro plants will be required to account for 33% of electricity supply by 2020.28 The EIA projects that even if support for renewables is Second. While these Two features of these technologies pose poten. which generally require utilities to obtain though power system flexibility will become specified percentages of energy from designated more important as the penetration of VERs renewable sources. they and some other technolo. The use of very long trans- more variable and difficult to predict than mission lines can cause technical problems and demand alone. In addition. increases. historically). Photovoltaic systems are more tolerant of clouds (diffuse light) and hence perform well across a broader swath of the U. VERs pose no new issues.S. require building more transmission than if as penetration increases.

The Federal matter.0 – 7. In many states.8 *Wind speeds are based on a Weibull k value of 2.31 sometimes.S. while interstate transmission of electricity was much less important than it is now or likely will be in the future.1 19.0 16.4 – 7.0 Source: This information was prepared by the National Renewable Energy Laboratory for the U. Cost allocation and siting have been particu.8 Location of U. Chapter 1: Challenges.8 5 Excellent 500 – 600 7.Figure 1.5 – 8. 2011. It is a combination of high resolution and low resolution datasets produced by NREL and other organizations. this process is larly contentious for transmission facilities that facilitated by the historic cooperation of states cross state borders or the 30% of U.7 4 Good 400 – 500 7. one or more federal agencies. gave the predecessor of FERC the right of eminent domain to site interstate natural gas pipelines.9 – 19. The data was screened to eliminate areas unlikely to be developed onshore due to land use or environmental issues.nrel. the special Power Act of 1935 made siting of all transmis.8 – 17. But as a general managed by federal agencies. the Natural Gas Act of 1938 as amended in 1947.0 – 8. Interstate pipelines were already important in the natural gas industry in the 1930s.7 – 16. In some regions of the country.3 – 15.9 6 Outstanding 600 – 800 8. able generation.5 15.7 – 24.renewable generation. the construction of to the efficient interstate transmission facilities requires the integration of renew. accessed November 16. 50 50 50 v In contrast to the Federal Power Act of 1935.S. Opportunities.pdf.gov/gis/pdfs/windsmodel4pub1-1-9base200904enh.8 – 11. and Major Recommendations 13 . Department of Energy.v Consequently.0 14. This image has been reprinted from the National Renewable Energy Laboratory’s GIS website http://www. Wind Power Classification Wind Resource Wind Power Wind Speed* Wind Speed* Power Potential Density at 50 m at 50 m at 50 m Class W/m2 m/s mph 3 Fair 300 – 400 6. difficulties of siting The special difficulties of siting sion lines a matter for the states rather than boundary-crossing boundary-crossing transmission FERC. consent of multiple state regulators and. land within an ISO’s territory.S. and lines that cross land managed by transmission facilities facilities will pose an obstacle federal agencies need the approval of those will pose an obstacle to the efficient integration of agencies.7 7 Superb 800 – 1600 8. Wind Resources United States – Wind Resource Map This map shows the annual average wind power estimates at a height of 50 meters.8 17. the wind resource on this map is visually enhanced to better show the distribution on ridge crests and other features.

October 20.3%. 2011. Department of Energy.9% per year between 2010 and 2030. particularly. allocating facility costs.5% in the 1973 to 2006 period. 2007) representing data from 1998–2005.nrel. electricity use grew at an average annual rate of 2. siting interstate transmis. 2008 Source: This information was prepared by the National Renewable Energy Laboratory for the U. which has resulted in an increasing ratio of system peak loads to average loads and falling capacity utilization.9 Location of U. vi Between 1949 and 1973. the EIA’s reference case projection is for growth to average only about 0. Author: Billy Roberts.S. disruption in the next few decades. several factors have contributed to this trend. Historically. This image has been reprinted from the National Renewable Energy Laboratory’s GIS website http://www. The data for Alaska are a 40 km dataset produced by the Climatological Solar Radiation Model (NREL. FINDING Electric Vehicles and Greater Demand Efficiently increasing the penetration of Variability grid-scale renewable generation while maintaining reliability will require modifica. Figure 1. to change in ways that pose challenges to the and. Concentrated Solar Power Resources Annual average direct normal solar resource data are shown. electricity use in the U.vi electricity In addition. The date for Hawaii and the 48 continuous states are 10 km satellite modeled dataset (SUNY/NREL. 2003).gov/gis/images/map_csp_national_lo-res. system.S. Even with rising prices after 1973. accessed November 16. demand has changed and is likely to continue sion system expansion. and the system was able to meet that demand with only sporadic difficulty. Although growth in electricity demand is not likely to emerge as an important source of tions to power system design and operation. In contrast. processes for planning transmis.S. The first of these has been a substantial sion facilities will need to be reformed.32 14 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . increase in power demand during select hours of the year. grew at an average annual rate of 8.jpg.

more than 20% of generation capacity (and.001 3.9 0.5%.001 5.10 also shows that. costs because of the need to pay for capital that is idle most of the time.001 6. since all new facilities must go in the 1980–1984 period in both New York and somebody’s backyard.0 0.001 7.001 2.7 Normalized Load 0. the fraction of U.000 hours so that more than 30% of capacity was for reliability. for instance.4 0.1 0.001 Hours of Year New England Average 1980–1984 NY Average 1980–1984 New England Average 2005–2009 NY Average 2005–2009 50 50 50 50 Chapter 1: Challenges. Because power systems need to be demand exceeded 70% of its peak for only about sized to meet peak loads with a reserve margin 1.2 0. Figure 1. it exacerbates This figure shows load duration curves expressed the need to build new generation plants and as percentages of peak hour demand to facili. the higher electricity rates must be. Between 1981 and 2001. demand exceeded 70% of its peak hottest summer days.10 Normalized Load Duration Curves for New England and New York 1.000 hours—about 11.0 1 1.001 8. in both New York and New England. market share of industrial use of electricity. the problem grew appreciably to 75.3% quarter century.8 0. By 2005–2009.33 This served to raise power demand worse.10 illustrates this change for New York the time.Figure 1. that in siting them.5 0.3 0. The costs of that idle capacity must Among the factors that have contributed to the be covered by ratepayers. in both New York and most substantially in the hottest hours of the New England. of the time. roughly. in the following homes with air-conditioning rose from 57. and Major Recommendations 15 . transmission lines and thus the problem of tate comparisons. We suspect this factor 30% of capacity was in use less than 12% of to have been particularly important in Figure 1. transmission and in use less than 12% of the time.6 0.4% By 2005–2009. Opportunities. distribution capacity) was in use less than 12% of the time. which generally corre- for only about 1.S. New England. and the more that increasing severity of this problem are the must be spent to build and maintain rarely spread of air-conditioning and the declining used assets.001 4. It shows. demand exceeded 80% of its peak for only about 1. Not only does this trend raise average and New England.000 hours so that more than spond to system peaks.

of the energy used in the U. particularly rooftop solar units. and their adverse effects on capacity sources. If they are charged when commuters tion. industrial of electricity demand and the future customers accounted for half of retail sales of penetration of electric vehicles will. the system saves only the wholesale cost return home. their impact which is typically recovered by per-kWh on the grid will depend on when they are charges. as seems most likely under current of energy.36 these vehicles cannot be compensate end users for generating their own ignored. tend to industrial load is. in turn industrial customers has declined steadily since will increase electricity costs. so the more important the absence of other changes. About 94% generators. they could add significantly to system end user saves both this wholesale cost and peak loads. When an end user increases genera- charged. they are deployed in large numbers. is dependence on imported oil. system comes from petroleum. sometimes. Thus net metering provides 16 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . however. in electricity. which include plug-in corporate tax incentives encourage distributed hybrids and battery electric vehicles.10. which are small-scale systems. demand when it would otherwise be low. and on average in the 2000 to 2009 period. they accounted for only 28% of retail sales. which current policies. As Chapter 5 perhaps the most visible of these. New York and New England in the period encouraged overnight charging could increase shown in Figure 1.39 The difference some high-income areas with environmentally between these rates is mainly the fixed cost of conscious consumers.34 Industrial plants often run around the clock all year.35 Increasing the penetration of EVs is tariff program for small-scale renewable often viewed as an attractive means of reducing generation. discusses. Whenever and wherever distribution (and. and these policies seem likely to utilization may be exacerbated by the spread of continue.S. transportation generally connected to distribution networks. the flatter the load duration accelerate the decline in capacity utilization curve tends to be. thus tending to flatten load duration curves. At the federal level.37 The California feed-in imported. A second factor that may have been more important in other regions is the relative FINDING decline in industrial use of electricity. worsening the problem depicted in the per-kWh charge used to recover fixed Figure 1. although their penetration generally is 46 states and the District of Columbia have projected to be slow at the national level under what are called “net metering” programs. however.38 In addition. the 1950s. This has made load duration Distributed Generation curves steeper. The relative importance of in the electric power system. Under net metering. personal and electric vehicles (EVs). measures that network costs. EVs are expected energy at the retail electricity rate rather than to achieve high levels of penetration quickly in the wholesale cost of energy. On Ongoing changes in the character average during 1950 to 1959. transmission). the policies. Existing policies at state and federal levels Looking ahead. This. On the other hand. and more than Most states have programs that subsidize half of crude oil and petroleum products are distributed generation.10. these trends are likely to favor distributed generation from low-carbon continue. even in the near term.

41 A 2008 complicate the ability to operate the distribu- tion system. Prompted by the vidual substations. to customer equipment. and sionals to learn new skills and knowledge. programs have languished over the past several decades due to the increasing popularity of other electrical engineering subdisciplines and a lack of research funding to support graduate students. causing Power & Energy Society of the Institute of unusual distribution flow patterns with power Electrical and Electronics Engineers (IEEE) flowing from the substation into the transmis. regulatory frameworks do not provide adequate indicated that workforce attrition could amount incentives for such investments. protection.40 In a widely cited April 2009 report.43 utilities have inadequate incentives to make Unfortunately. distributed generation can workshop on this topic in November 2007. distributed to rejuvenate its workforce in order to maintain generation simply reduces the load at indi. current levels of performance. subdisciplines and a lack of research funding to Enabling such penetration in a cost-effective support graduate students. power and energy however. the exceed load at the substation level.44 Furthermore. high-voltage swings. Opportunities. and reasons in the subsequent five years. and control will be necessary to enable signifi- increasing popularity of other electrical engineering cant penetration of distributed generation. to 40%–50% by 2013 across a wide range of power industry technical job categories. Even if it faced none of the challenges discussed above.S. and technicians.S. plant operators. while distributed survey conducted by the Center for Energy generation would reduce its sales.S. Power and Energy Engineering sion grid. university power engineering investments necessary to accommodate it. Additional monitoring languished over the past several decades due to the and new standards for operation. results of a National Science Foundation tion. meeting the challenges complicates the design and operation of and realizing the opportunities discussed in this distribution systems. an industry consortium. Net metering provides study will also require many industry profes- a subsidy to distributed generation. which can be detrimental the PWC noted that approximately 45% of U. this condition can produce workforce. to strengthen the U. a recent survey Chapter 1: Challenges. and Major Recommendations 17 . In fact. the electric power industry would need At low levels of penetration. At high levels of penetra.42 Beyond retirements. however. manner would require incremental investment by the distribution utility.an additional subsidy to distributed generation Aging Workforce of all sorts that may encourage uneconomic penetration. The systems involved are currently Workforce Collaborative (PWC) to lead efforts not designed to handle such reverse flows. pipefitters and pipelayers. particularly during emergencies University power engineering programs have and planned outages. such as circuit breakers. electric utility engineers would be eligible for tion can also add to the stress on electrical retirement or could leave engineering for other equipment. High levels of penetra. Current Workforce Development. High penetration of distributed generation engineers. founded the U. FINDING including lineworkers.

FINDING As discussed in Chapter 2. multiple customers have reported the outage. and 27% are expected to periodic equipment failures. the PWC has of the transmission system and thus published a detailed action plan with recom- mendations for a wide range of industry make possible more efficient and stakeholders.S. such as the loss of a generator demands and both the pipeline of students or transmission line. many utilities workforce. gencies.46 The DOE also recently awarded reliable operation. 18 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . for instance. time-stamped widely recognized.45 While it is difficult to predict exactly rapidly disconnecting lines or generators where how many new engineers will be needed problems occur. innovative technolo- Because of its aging workforce and the gies can improve operator knowledge about the state of the transmission system and thus make nature of emerging challenges. it remains to be seen data on transmission system conditions that whether efforts to deal with it will prove system operators can use to anticipate contin- adequate. and customers are often unaware when workers and develop a variety of power system problems occur on the bulk power system. PMUs are being widely deployed.47 These projects are events. however. enhance system efficiency. While this problem has been provide rich streams of frequent. If a tree limb takes down ensure the availability of a qualified technical a distribution line. indicated that approximately 40% of power Technologies for Reliability and Efficiency engineering faculty members at U. system operators maintain entering power engineering and the faculty a prescribed level of generation reserves and in place to train them. To guard against the negative between now and 2030. industry workforce challenges have Innovative technologies can improve received increasing attention in the past several operator knowledge about the state years.000 time. and the issue will likely do not know where to send a repair truck until continue to receive attention in the years ahead. Among other efforts. phasor measure- of skilled workers. universi- ties will become eligible for retirement within The electric power system is built to handle the next five years. utility industry faces a near-term shortage On the transmission system. reduce the risk of wide-area blackouts. These measures work well most of the expected to help train as many as 30. education programs. $100 million to 52 workforce training and development efforts. With funding from the American Recovery and Reinvestment Act of 2009. It is too early to tell Customers more often observe failures of the whether these efforts will be sufficient to distribution system. updated procedures for reacting to unexpected Fortunately. there appears to be a consequences of contingencies in the bulk significant gap between anticipated industry power system. convert data from these systems to actionable informa- tion. and improve system models. primarily by retire. particularly power ment units (PMUs) are powerful devices that engineers. the electric possible more efficient and reliable operation. but work is needed to network these devices into systems. and employ this information in the control of the grid.

electricity demand more sensitive to system sive monitoring and control equipment. enhance distributions systems and make demand more responsive to real-time costs. especially at the retail level. This conditions will require changes in regulatory is cost-effective. flexible alternating each system is different. in recent years U. Given the many new chal- lenges that will confront distribution systems in the next 20 years. a variety of new and emerging technologies. to which customer each distribution line serves fewer customers response. resulting in more efficient use of grid assets and thus lower rates. Higher possible to provide real-time incentives to penetration of VERs is likely to increase the reduce system peaks caused by central air- value of deploying these technologies in the conditioning. however. is poorly than a transmission line. and cost is always an current transmission system (FACTS) devices important consideration in the choice of based on advances in power electronics can technology. transmission system. can receive price more power to be transmitted on existing lines information based on the real-time cost of without increasing the risk of failure. and tively novel pricing regimes. As Chapter 7 notes. some modernization and enhancement will surely be appropriate. utilities and regulators have tended to avoid investments in unfamiliar technologies perceived to have uncertain payoffs. Compared to the transmission network. even if no customers call. including FINDING distribution management system software. as there are far more miles of policy to encourage the deployment of rela- distribution lines than transmission lines. AMI. but Chapter 1: Challenges. many available technologies have not yet been widely implemented at the distribution level. transmission level than the distribution level. but providing electricity and can transmit usage historically the incremental benefits generally information every few minutes. will enable system operators to operators’ ability to observe and control detect an outage and identify its cause within these systems. Moreover. Opportunities. and other loads.In addition to PMUs. the the effective use of these technologies to make distribution system uses more basic. inexpen. There are many understood. for instance. In addition. Because it is often more cost-effective to invest but effective use of these technologies will in monitoring and control systems at the require changes in regulatory policy.S. including FACTS and other new technologies can allow advanced metering systems. Technologies similarly can minutes. and Major Recommendations 19 . as Chapter 6 explains. automatic to improve the reliability and efficiency reconfiguration of distribution circuits. and of bulk power systems by enhancing advanced metering infrastructure (AMI). vehicle charging. This makes it have not justified the associated costs. New technologies have the potential more accurate control of voltages. as discussed in Chapter 6. provide greater control of voltages and power flows throughout the bulk power system. technologies to enhance the distribution system.

These websites include or unnecessary barriers to adoption of smart SmartGrid. Deployment and integration of advanced to a wide variety of electric grid modernization electricity storage and peak-shaving tech- efforts and ideas over the past several years. and IEEE’s Smart definitions of the smart grid. BOX 1. practices. including renewable resources. recent or anticipated technical innovations can 5. Smart Grid Information Clearinghouse The scope of this study is broader than many (www. concepts. and thermal-storage government. some of which seem only loosely con. and services. we do Grid site (http://smartgrid. the grid technologies. including plug-in electric and While uses of the term vary throughout industry. Deployment of `smart’ technologies (real-time. and energy new services. it is perhaps best airconditioning.org/). and resources and generation. The term “smart grid” has been used to refer 7. Title XIII of the Energy Independence smart grid discussions. applications that are enabled by the SEC.ieee. 1. Dynamic optimization of grid operations on the term in this report. ture investments are discussed in Chapter 7. For example. sion and distribution system to maintain a reliable Chapter 9 analyzes many of the challenges and secure electricity infrastructure that can meet related to expanded data communications use. We have focused and resources. instead on the broad goal of making the grid of 3.gov (www. measurement units. Integration of `smart’ appliances and consum- er devices. throughout distribution networks are described It is the policy of the United States to support the in Chapter 6. Development and incorporation of demand and. products originating from industry. and efficiency of the electric grid. Deployment and integration of distributed the future more resilient. nologies. STATEMENT OF POLICY ON deployment of sensors and communications MODERNIZATION OF ELECTRICITY GRID. Increased use of digital information and Because the term “smart grid” means different controls technology to improve reliability. devices that have the opment of the smart grid as national policy and potential to equip bulk system operators with identified it as a broad collection of ambitious greater real-time knowledge of the state of the goals. future demand growth and to achieve each of the including cybersecurity and information privacy following. and control systems throughout and control options. sgiclearinghouse. consider the technologies that are core to most In the U. and advanced metering infrastruc- modernization of the Nation’s electricity transmis. organiza- 10. Identification and lowering of unreasonable tions. bulk transmission system.smartgrid.2 THE “SMART GRID” 6. further these goals.org/). interactive technologies that optimize the physical operation of appliances and consumer devices) for metering. Seizing opportunities related to efficiency resources. and the public. communi- cations concerning grid operations and status. reliable amid a variety of emerging challenges 4. meanings are evolving. as follows: Chapter 2. all levels of the electric grid. secure. demand-side resources. are discussed in nected to the grid’s intelligence. and distribution automation.gov). 20 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .S. with full cybersecurity. Development of standards for communication websites have been created to try to make sense and interoperability of appliances and of the flood of “smart grid” ideas. perhaps. things to different people and because its security. efficient. and individuals. described as the expanded use of new communi. However. 8. hybrid electric vehicles. which together characterize a Smart Grid: challenges. phasor and Security Act of 2007 established the devel. automated. to enable the provision of desirable response. and equipment connected to the electric grid. we have avoided reliance 2. including the infrastructure serving the grid. 1301.. Provision to consumers of timely information cations. sensing. Many industry 9.

increasing amounts design and operation of grid communications of data will be exchanged among meters. Since no communications system satisfy the concerns of customers and their can be completely free from errors. when later generations of equipment protect the privacy of grid information to are installed. personal data and ensuring consumers’ privacy will be an important consideration in the As explained in Chapter 9. Utilities and related organizations various components to interoperate now and in will have to develop systems and procedures to the future. standardization for interoperability and The existence of more communications nodes and channels facilitates the placement of of cybersecurity and raises serious issues MALICIOUS DATA INTO THE SYSTEM IN ADDITION. facilities through a complex communications although not yet specifically electricity usage system that must follow standards that allow information. local power of the electric grid itself will soon be available systems enhanced reliability overall but made at a level of detail that will be of value to those possible wide-area blackouts. increasing use of new sensing and automated response technologies will enhance reliability Deciding who has access rights to these and efficiency overall but create new problems. Information about the operation The interconnection of small. nor by today’s policy makers. the with both commercial and malicious interests.Cybersecurity and Privacy envisioned by today’s utilities. More chilling is the FINDING possibility of deliberate sabotage via computers Greater reliance on data communications and data communications. Similarly. other networks. Many governments have passed laws sensors. and various computers and control protecting the privacy of personal information. the sort of cyber- in the grid increases the importance of attacks that other industries have experienced. grid must be designed to mitigate the conse- quences of data errors. the future governments.

transmission. infor. Two broad points deserve mention here. First. of privacy. and sharing are collected in the final subsec- and storage of increasing amounts of infor.3 MAJOR RECOMMENDATIONS damage. and Major Recommendations 21 . Opportunities. mation also comes heightened concern for protecting the privacy of that information. and cybersecurity. This section highlights what we consider the With the collection. the adoption of coherent and stable national mation on personal habits will be available policies on greenhouse gas emissions. tions below. and renewable and distributed Chapter 1: Challenges. electric to electric companies at a level never before vehicles. pricing and regulation of distribution. As advanced metering is implemented. greater reliance on automated responses to system conditions that may be misreported can make it more difficult to prevent serious 1. Major recommendations for research and analysis and for data development With the collection. processing. and storage of increasing chapters and discusses some of our related amounts of information also comes findings and conclusions. transmission. We present recom- heightened concern for protecting the mended policy changes affecting transmission. privacy of that information. most important elements of the more than 20 specific recommendations developed in later processing.

should significantly increase wide-area planning of transmission systems Support for organized wholesale electricity markets and rationalize the allocation of the costs of along with open. purposes unrelated to energy manage 30% OF 53 LANDS THEY CAN AND DO EXERCISE THE As the electric power system has become power to block or delay the construction of increasingly interconnected. costs should be allocated As generation from wind and solar power as closely as practical in proportion to antici- grows in importance. We nonetheless believe the public transmission grid has been repeatedly affirmed interest would be served if the affected parties as national policy. System operators here Under current law. as we noted earlier. renewable generation are likely to accelerate tives and thereby accelerate the appropriate this trend because many of the best wind and evolution of the grid. interests sometimes conflict. interregional projects. issued in July 2011. and their interests developed. as would requiring new VER Under current law. nondiscrimination is still a work in progress. Moreover. and their appropriate for operation in a high-VER future. support for solar resources are far from major load centers. but development and basis. public policies favoring 22 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . FERC Order No. respond to substantial changes in the output of these variable resources. Similarly. mechanisms to ensure can block a multistate project. went beyond the order’s planning requirements and established permanent and collaborative These policies are of central importance for the processes for transmission planning at the transmission grid. that power systems are adequately flexible will federal agencies with missions that include become more important in the future. Any involved state And. In that procedure. organized wholesale electricity markets along with open. states retain the primary generators to meet performance specifications role in siting transmission facilities. and solutions are being facilities. as discussed earlier. especially. in Europe are aware of the primary role in siting transmission problems involved. nondiscriminatory access to the Transmission system expansion in many transmission grid has been repeatedly affirmed regions is routinely planned on a multistate as national policy. And. nondiscriminatory access to the transmission facilities that cross regional boundaries. consolidation of small balancing areas would be helpful. and attempts at interconnection-wide expansion of the reach of open access and planning are under way. Second. power systems will have pated benefits and other efficiency-enhancing to become more flexible to be better able to principles should be followed. and organized electricity interconnection level that combine top-down markets with many common market design and bottom-up approaches. as Chapter 4 explains. we believe all affected parties should develop a single procedure for Bulk Power and Transmission Systems each interconnection. states retain the and. It seems clear that real or virtual sometimes conflict. the importance transmission lines across these lands in cases of transmission lines that cross state borders of perceived conflict with other land manage- or federal lands has also increased. generation would enhance investment incen. while elements will continue to define the framework Order 1000 only requires the development within which they can be effectively of bilateral cost-allocation procedures for implemented. ment missions. 1000.

Existing Renewable Energy Rapid Response Task Force. recog- nizing the growing importance of interstate As noted earlier. In recognition Since an essential requirement of a power of the increasing importance of interstate system is the ability to meet peak demand. in recent decades the ratio of natural gas pipelines. electricity transmission. The penetration of the event states disapproved construction of electric vehicles may exacerbate this trend multistate electricity transmission facilities. if not the default. systems and residential appliances. time intervals to reflect the often dramatic changes in the actual cost of providing elec- Some have argued that in the interest of tricity—combined with technology to auto- efficiency. now including increased. Substantial AMI investments have recently been funded through the Recovery Act of 2009. and some state regulators have Chapter 1: Challenges. similarly pation in some siting decisions by creating the responsive to system conditions. While both approaches clearly have strengths and weaknesses. FERC should have sole siting mate response to price changes. and Major Recommendations 23 . the power of eminent domain. boundaries or federal lands (Chapter 4). As unless their owners can be induced to charge Chapter 4 discusses. ultimately. Congress gave FERC peak electricity demand to average demand has authority to site these facilities. However. the Energy Policy Act these developments have reduced capacity of 2005 contained a section that was intended utilization and thereby increased average cost to give FERC useful backup siting authority in and. with third parties generally enabled to compete to provide equipment to automate enhanced siting authority for major response to price changes. Even greater savings decisions have effectively annulled that section. however. subsequent court them in off-peak periods. authority over major projects. While the Obama administration has recently including commercial and industrial HVAC taken steps to streamline federal agency partici. and residential dynamic pricing requires substantial invest- ment in AMI to measure usage over short time intervals. legitimate concerns. in part because the behavior of residential consumers faced with dynamic pricing is not yet adequately understood. retail rates. and this trend is likely to continue. as it does over interstate natural gas pipelines.The federal government has addressed this type Pricing and Distribution Regulation of structural problem before. adopting Many large commercial and industrial either would be a significant and important customers now operate under dynamic pricing. may be realized by making other loads. studies suggest that regulators and utilities can the current system for siting transmission achieve this using a combination of dynamic facilities remains a significant barrier to effi. for residential R E CO M M E N D AT I O N consumers also by the end of our study period New legislation should grant FERC in 2030. Others contend The behavior of residential consumers faced with that giving FERC effective backstop authority would create a process more sensitive to states’ dynamic pricing is not yet adequately understood. response transmission facilities that cross state automation technologies are not yet mature. In 1938. Opportunities. pricing—in which retail prices vary over short cient expansion of the grid. We believe such pricing regimes will be wide- spread options. improvement over the status quo.

R E CO M M E N D AT I O N utilities that have committed to AMI State regulators and those who supervise systems should begin a transition to government-owned and cooperative dynamic pricing for all customers and utilities should recover fixed network costs publicly share data from their experiences primarily through customer charges that (Chapter 7). Utilities that have not committed to AMI systems and for which the operational benefits of these systems are less than their cost should These fixed charges could depend on indicators take advantage of the option to learn from early of customers’ need for network capacity. For adopters before making a decision to invest. regulators and utilities should modification of the distribution system. the same regardless of the energy source—clean mandated investments in AMI have provided solar or dirty diesel—used by the distributed to develop efficient paths to universal dynamic generator. R E CO M M E N D AT I O N With the cooperation of their regulators. challenges are real. and transmission and distribution networks demand response in coming years will require through volumetric charges per kilowatt hour significant investments in new and emerging of use. For example. they do not appear to be tion charge for that electricity. deal with the growth of distributed generation and energy efficiency initiatives. But saves only the corresponding generation cost movement toward the dynamic pricing regimes because the cost of distribution is almost that AMI enables has been slow. that exploit the important learning opportunity fixed cost may be increased. While the purchasing it from the local distribution utility technical problems associated with these new saves both the energy charge and the distribu. The remedy is straightforward. electric vehicles. contribute more to local peak demand based on effective competition in the retail sales of their pattern of prior consumption could pay a electricity may stimulate innovation in ways higher fixed charge than customer groups that to make dynamic pricing both acceptable to are expected to contribute less.) This outcome is the Recovery Act-supported and regulator. Utilities Coping efficiently with the integration of currently recover the largely fixed costs of distributed generation. However. if a high concen- enormous potential value of dynamic pricing tration of distributed generation required of electricity. mandated universal AMI deployment. customer groups that are expected to jurisdictions with wholesale electricity markets. may differ among customers but should not vary with kilowatt-hour consumption (Chapter 8). least in principle: recover fixed network costs mainly through nonvolumetric charges. at pricing—and then to follow those paths. charges for cost recovery should improve utility incentives by decoupling utility revenues from Electricity pricing also needs to be changed to short-run changes in sales. a customer who will aim to provide new capabilities. (Indeed. Under this regime. not just generates electricity on-site rather than expand capacity in traditional ways. Systems that consumers (and regulators) and effective in continue to rely significantly on volumetric modifying demand. Given the entirely unchanged. the tendency of traditional 24 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . but the utility serious. a practice that distorts the relative prices technologies that will be riskier than most of central station generation and distributed RECENT INVESTMENTS IN DISTRIBUTION SYSTEMS THEY generation.

this is an As communications systems expand into every important problem—but one without an facet of grid control and operations. their complexity obvious solution. while we expect that opportunities for response. responsibility for overseeing grid cybersecurity most likely. In addition. since both regulators and and continuous evolution will preclude perfect utilities seem to be punished for bad outcomes but not rewarded for good ones. issues. The National across all aspects of grid operations. protection from cyberattacks. the appropriate regulatory authority to Resolution of the former debate rests with the enhance cybersecurity preparedness. Nonetheless. so will problems of deserve serious attention. As Chapter 8 discusses. Response and recovery. their and more expensive over time if increasingly complexity and continuous evolution will attractive opportunities to enhance efficiency preclude perfect protection from cyberattacks. State PUCs (which are generally responsible only for investor-owned distribution investments in unfamiliar technologies systems). there a single agency to have responsibility are ongoing debates about the use of spectrum for working with industry and to have and the roles of public and private networks. As data communications becomes more and distribution-level cybersecurity risks important in the grid. while also ensuring that the returns and other public systems generally lack the on these investments are shared expertise necessary to deal with cybersecurity appropriately with ratepayers. cooperatives. Opportunities. and distribution systems (Chapter 9). municipal electric systems. and reduce cost through innovation are not exploited. regulatory innovations are necessary to provide adequate incentives for investments in unfa. Institute of Standards and Technology (NIST) is overseeing the complicated process of developing the relevant interoperability stan. cybersecurity. and Privacy level. While the consequences of a successful attack on the bulk power system are potentially much greater than an attack at the distribution Communications.regulatory systems to encourage excessively As communications systems expand into conservative behavior is likely to become more every facet of grid control and operations. and Major Recommendations 25 . Thus no organization currently has components over a variety of networks with. in addition to prepared- miliar technologies while also ensuring that the ness. including both bulk power unequally. NERC is responsible for standards priately with ratepayers. NIST is facilitating cybersecurity and issues of privacy. Cybersecurity. will thus be important components of returns on these investments are shared appro. FCC. R E CO M M E N D AT I O N dards. As the grid the development of cybersecurity standards evolves. development and compliance for the bulk power system. This process is critical. and recovery across the electric both public and private networks will exist unless the regulatory environment treats them power sector. Chapter 1: Challenges. the boundary between transmission and distribution has become increasingly blurred. but it does not have an operational ability of different types and generations of role. a variety of owners. but no entity has comparable Regulatory innovations are necessary nationwide responsibility for distribution to provide adequate incentives for systems. and it should be The federal government should designate encouraged and supported. it will be critical to maintain interoper- broadly.

expert organizations as IEEE and EPRI. For this to happen. and legislative Currently. to be actively debated in several states. this investment provides an important learning Coordination across states will be necessary to opportunity. the electric utility industry traditionally has relied primarily on its uncertainty for networks with suppliers for the innovation that has driven its anything approaching the complexity productivity growth. and regulators. public power authorities. and PUCs. the issues involved in the use and systems are required to integrate PMUs and protection of customer electric usage data are FACTS devices effectively into system opera- complex. They include cannot do multiperiod optimization under the development of computational tools and uncertainty for networks with anything well-designed social science–based studies of approaching the complexity of the Eastern or customer response to dynamic pricing regimes. of PMU penetration in the grid. and communication Finally. particularly because of the many tions. Supplier R&D naturally of the Eastern or even the Western has focused on equipment that can be sold to Interconnection. one or other relevant federal agencies. Research and Analysis Existing planning methods cannot do multiperiod optimization under As noted earlier. It concerns. New algorithms. utilities. and these are not likely to attract problem becomes. even the Western Interconnection. Existing planning methods traditional equipment vendors. Additional efforts in several non- equipment related research areas relevant to The wider the area over which transmission tomorrow’s grid are likely to have substantial planning is done. electric utilities generally lack both proposals designating either a combination of appropriate financial incentives and the exper- FERC and DOE or the Department of tise necessary to perform either type of research Homeland Security (DHS) have been advanced. establish baselines for future operational tools as data on both companies and their customers that can monitor and control networks with cross state boundaries. it to reverse the downward trend in cooperative should take all necessary steps to ensure that R&D spending and make appropriate use of it has appropriate expertise by working with cooperative funding through EPRI. software. The Recovery Act has funded expansion different and evolving views of consumers. the more complex the payoffs. will also likely be necessary for the industry Once a lead agency has been designated. and such project-specific coalitions. but the industry should The various agencies each have both strengths nonetheless be able to support the modest but AND WEAKNESSES FOR SUCH A DESIGNATION THE sustained efforts required. underscoring the need for action. state more independent system operators. This may require new legislation. Like the utilities. If shared. because of the challenge of efficiently 26 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . but ongoing progress benefits consumers broadly and permit jurisdictional confusion raises security modest increases in utility R&D budgets. data generated by existing mitigate concerns of companies that operate in PMUs can be used to develop algorithms and multiple jurisdictions and of their customers. earlier. greater PMU and FACTS penetration. the grid is becoming more closely that have been made possible by recent AMI coupled at the interconnection level partially investments. As noted perhaps supported by response automation. NERC. and development. These issues continue Recovery Act–financed investment in AMIs. capabilities of any agency can be enhanced regulators will need to recognize that technical to address its relevant weaknesses.

The electric power industry should fund We have identified three additional areas in additional research and demonstration which ensuring the appropriate availability projects to develop: computational tools of useful data would be particularly valuable. have a good deal of experience making Chapter 1: Challenges. a year and a half later. data are not shared as widely as would be cies (i. the benefits of PMUs may must be designed and tested in the field. the general availability of detailed and understanding of consumer response data on the U. There will be repeatedly. If this initiative fails to achieve the consumer engagement and education strategies hoped-for results. as discussed earlier.e. and effective ments. and FERC’s In the course of this project. Finally. The responsible federal agencies should take steps to ensure that the critical data R E CO M M E N D AT I O N at issue are shared appropriately. a successful attack at some point. NERC created two nondis- these systems and response automation tech. NIST. to support analysis. but federal agencies systems (Chapter 7). closure agreements to facilitate sharing PMU nologies to make electricity demand more data. As Data Development and Sharing discussed earlier. Among other many utilities with significant PMU deploy- things. not be realized. the industry should One promising recent initiative has been use the first round of AMI deployments to undertaken to enhance high-value data sharing. for the more complex Eastern Interconnection. and industry. Sometimes potentially valuable data working with the private sector in a coordi. some interconnection-level planning efforts are under way. ingly find themselves. bulk power system would raise to alternative pricing/response automation serious security concerns.S.. especially in the unfamiliar situations in and distribution systems. it persists. though this problem We have been struck repeatedly has been observed and by shortcomings in the data As noted earlier.S. such data are not available methods for wide-area transmission plan.48 available on the U. and DHS). Unfortunately. further research on consumer reactions ment activity had not yet signed these agree- to dynamic pricing is needed. Good data are to develop best practices for response to and critical inputs to good decisions regarding the recovery from cyberattacks on transmission grid.S. to support the research necessary ways that limit their usefulness. electric grid. FERC. and to decision makers in both government discussed in Chapter 4. that will exploit the potential of new hard.integrating remote renewable generators. and recovery from cyberattacks (Chapter 9). beneficial. so that such practices which public and private actors will increas- can be widely deployed. Opportunities. 1000 calls for an expansion of the struck repeatedly by shortcomings in the data geographic scope of planning processes. It is thus Sometimes available important for the involved government agen. learn how best to employ the capabilities of In February 2010. and Major Recommendations 27 . electric grid to researchers The development of new planning methods. we have been Order No. While network data on the Western Intercon- ware to improve monitoring and control nection are available at a level of detail adequate of the bulk power system (Chapter 2). Obviously. Even tial payoff. ning (Chapter 4). DOE. responsive to system conditions. are simply not collected or are compiled in nated fashion. perfect protection from commented upon cyberattacks is not possible. available on the U. processes for response to This inhibits both wide-area planning and the improvement of wide-area planning methods. thus has a high poten.

As results become 1. As this study indicates. and other dimensions of shared widely. many jurisdictions do not even participants in electricity markets (including require utilities to report data on reliability consumers) with policy goals. concern and. of comprehensive. If the grid is to evolve used to power computers and other complex with minimal disruption despite the challenges electronic equipment rises. performance (Chapter 8). especially those in different jurisdic- But the journey to the electric grid of 2030 has tions.4 CONCLUDING REMARKS available. Several websites. The environment in which performance as well as cost. 28 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . Achieving the tors and others supervising distribution full potential of these distribution system utilities should require utilities to compile technology demonstration projects will require and publish standardized metrics of utility that data on both successes and failures are cost. confront significant new challenges and inevi- tably undergo major changes. But.gov and www. there is no crisis here. action. comparable data hinders regulators’ attempts to evaluate utilities over We are encouraged by recent levels of aware- time or make useful comparisons across ness. development and publication of along the way. While regulators ahead and if electricity rates and levels of have at times considered performance in an ad reliability are to be acceptable. a more systematic approach to focus on meeting the system’s challenges. it is imperative that resources such as these are effectively used to share data and Between now and 2030. let alone data on efficiency needs to conduct research in key areas and both or other aspects of performance. and make it easier for regulators to provide meaningful incentives for good performance. case-specific fashion when setting allow- in government and industry need to continue able rates of return. rhetoric. confidential data available for use in research R E CO M M E N D AT I O N while ensuring that confidentiality is not FERC should require that detailed data on violated and data are not copied. utilities. To the ultimate benefit of all stake- begun. Regulators should seek to incentives based on performance metrics. A range of system-level issues need to be Some U. reliability. would almost certainly produce better results. State regula- Investment Grants initiatives. Despite alarmist Electric utility customers care about reliability. regulators and many abroad have addressed. develop policies that better align incentives of as noted earlier. the electric grid will lessons learned. have been established to disseminate informa- tion about these projects. formal used as appropriate.S.S. and those concerns the grid will operate will change substantially are arguably increasing as the share of energy in the next two decades. and new technologies need to be accordingly begun to establish explicit. and other dimensions of advise complacency. including www. and there will be plenty of surprises holders. the U.org. smartgrid. The industry in a useful form. The lack collect and share important data. bulk power system be compiled and made appropriately available (Chapter 4). in some key areas. In this regard. much can standardized cost and performance metrics and should be done now to smooth the poten- would facilitate assessment of utility outcomes tially very bumpy road ahead. decision makers hoc.sgiclearinghouse. But we do not customer service. we note that there is a lack of DOE should work to ensure that compre- information currently available on projects that had been funded through DOE’s Smart Grid hensive data from its Smart Grid projects Demonstration Program and Smart Grid are widely shared (Chapter 6).

Transformed Our Lives (Joseph Henry Press.S.html.com/content/defg/abaccus.S. Constable and B. Energy Information Administration.” http://www.” Financial Times. Electric Power Industry Overview 2007. Department of Energy.S. doe. 2010). “Electric 15 2 U. http://www. Revenue.asp. Somerville. Energy Information Administration.gov/cneaf/electricity/page/prim2/toc2. DC: U. DC. “Texas Declares New ‘Energy 16 http://www. 2011.com/cms/s/0/430f3f08-be89-11e0- 3 U. 2010). August 4.S. Sales.ft. Emergency’. Innovation: Twenty Engineering Achievements That defgllc. 2003). http://www. U.REFERENCES Distributed Energy Financial Group LLC.gov/cneaf/electricity/esr/table5_a. Annual 14 Baseline Assessment of Choice in Canada and the 1 G. Energy Information Administration. and Average Price 2009 (Washington. A Century of United States (Washington.eia.eia. Table 5A. S. McNulty. AB FEABDCHTML " 6ASTAG. html.

The Washington Post. h/BAMA Coordinated Bulk Power Supply Program Unveils Plan to Speed Smart-Grid Development. DC: U. Department of Energy.” (Washington.S.

*UNE .

.

Amin. 12 U. W. “Electricity Wholesale Market Design Council of European Energy Regulators. H. LaCommare. http://www. DC: U. 4 U. see note 3 J. DC. F. Reliability of the U.S. M. see note 17 19 RFF Press. note 8 above. 2011). E. 10 North American Electric Reliability Corporation. G. K.php?cid=4|38.nerc. S. 21 “Understanding Bulk Power Reliability: The 9 E. Pollitt.S.S. “Liberalisation and R&D 22 11 S. Fisher. B. H. Giles and K. Kammen. Environmental Protection Agency. 2008).” IEEE Spectrum. Distribution 17 (Washington. State Increasing Need and the Feasibility of Expansion. Energy Information Administration.S. eds. and R. Energy Research and Development: Declining Investment. “U. Greenhouse Gas Emissions and Sinks.” Electricity Profiles 2008 (Washington. H. 6 U. January 2011. 2008). see note 2 An Assessment of Publicly Available Information above. Inventory of U. ! AND 2010). Energy Policy    n AND 6 2EZENDES. (Palo Alto.S. Moselle. LaCommare. J. Energy Information Administration.gov/ Reliability Indices Tracking within the United States climatechange/emissions/usinventoryreport. International Conference on System Science. DC: Electric Power Research Institute. H. Electric Power Transmission: in Network Industries: The Case of the Electricity Background and Policy Issues (Washington. 2009 UDI Importance of Good Data and a Critical Review of Directory of Electric Power Producers and Existing Sources. Jamasb and M. Eto.” Research Policy    n Congressional Research Service. 5 W.org/Xplore. Belgium.” in Harnessing Renewable Benchmarking Report on the Quality of Electricity Energy in Electric Power Systems. Schmalensee (Washington. Reported to State Public Utility Commissions (Berkeley.S. 2009): 27. Eto and K. CA.” Proceedings of the 45th Hawaii Distributors (New York: Platts. Supply 2008 (Brussels.1990–2009 Electric Power Research Institute. January 2012. eds. Brown.S. Kaplan.epa. Hogan. 2010).html. http://www.com/page. http://ieeexplore. 2008). Tracking the 20 above.ieee. Padilla.. M. 4th 18 in a Low Carbon Future. Energy Information Administration. Electric Power System: 7 U. and K. Electricity Supply and Demand Database (April 2009). T. above. Electrical Grid Gets Less Reliable. J. DC: Industry. CA: Lawrence Berkeley National 8 North American Electric Reliability Corporation Laboratory.S.S. “U. 2003). Nemet and D.

doe. DC: U. R&D (Washington. Chapter 1: Challenges. lpc/iprprodydata.htm.gov/todayinenergy/ Energy in Electric Power Systems.” Today in Energy. Energy Information Administration. J. Moselle. Department of Energy. ed. http://www.pdf. Statistics (February 25.S.eia. detail.cfm?id=1450. Padilla. DC: RFF Press. in the United States.bls.gov/todayinenergy/detail. 2010). 13 U. Opportunities. and R. General Accounting Office. 2009). 2011. “Electricity Retail Choice Is Mandated in Texas and U. Schmalensee. “Renewable Electricity Generation 24 Electricity Retail Choice States.S. CFMID AND 53 %NERGY )NFORMATION Administration. Bureau of Labor Statistics Productivity 23 Growing in Three States. 1998).gov/ 2011. B. Schmalensee (Washington. http://www. http://www. and Major Recommendations 29 . “Participation Lags in Most R.eia.” Today in Energy.S. 209–232.eia.” in Harnessing Renewable May 19. Electric Utility Restructuring: Implications for Utility gov/cneaf/electricity/st_profiles/sep2008. http://www. May 18. 2010).

dsireusa.org/ 30 European Network of Transmission System images/pdf/US_Power_&_Energy_Collaborative_ Operators for Electricity. DC.” http://www. et al. W.html.ieee-pes. 26 U. Renewable Generation.org/. “California A Collaborative Effort to Strengthen the U.htm. 25 North Carolina Solar Center and the Interstate California Public Utilities Commission..S. Renewable Energy Statistics & Data. Reder. Energy Information Administration. http://www. Foundation for 29 U.gov/ http://www.gov/cneaf/electricity/epm/ above.cpuc. http://www.eia.” http://www. Research and Action_Plan_April_2009_Adobe72. Renewable Energy Council (IREC).gov/renewables/index.S. “Engineering the Future: 40 27 California Energy Commission. Power and Energy Engineering Workforce 41 Collaborative. http://www. Annual Future Electric Energy Systems: A Strong Power and Energy Outlook. Database (IREC) of Tariffs Available for the Purchase of Eligible Small State Incentives for Renewables & Efficiency.ca. Electric North Carolina Solar Center and the Interstate 39 Power Monthly September 2011 (Washington. 28 Ibid. Energy Magazine 8 (July/August 2010): 27–35.eia. DC. Energy Society.ca. “Feed-in 38 Renewable Energy Council.S.” IEEE Power & energyalmanac. No-Sunset Case (Washington. Gaps 42 Challenges and Beyond "RUSSELS.pdf. see note 25 2011). Development Plan: European Grid Towards 2020 Center for Energy Workforce Development. Energy Engineering Workforce (IEEE Power & 2011). 2009). epm_sum.S. U.gov/oiaf/aeo/tablebrowser/.S.html. Preparing the U. Energy Information Administration. PUC/energy/Renewables/hot/feedintariffs. Power and Energy Workforce.

"ELGIUM.

” presented at the 2010 IEEE Power & Energy 31 A. “Focus on Education—‘Smart’ Electric 43 IVGTF_Outline_Report_040708. NJ.cewd.pdf. “Two Critical Barriers to Transmission Society General Meeting. “Accommodating High Levels of www. MN. 2008). http:// Corporation.nerc. http://www. 2009. at the TAPS Conference. Portland. B. Variable Generation.” white paper (Princeton.   in the Energy Workforce Pipeline: 2008 CEWD and North American Electric Reliability Survey Results (Washington. Brown. “Power Education at the 44 Crossroads.” presented July 25–29.org/documents/CEWD_08Results.pdf. Minneapolis. ME. Development: Siting & Cost Allocation. Albuyeh. Power Systems 101: An Employer’s Perspective. Chowdhury. 2009).” IEEE Spectrum. 2010. DC. October 19.com/docs/pc/ivgtf/ F.

/CTOBER .

Department of Energy. Census Bureau.gov/prod/2005pubs/ for the Electric Power Sector. Electricity Transmission in a Restructured Industry: doe..S.S. of Energy. 35 Ibid.gov/totalenergy/data/annual/index. DC: U. and Poverty.census. DC: U.S. technology-development/smart-grid/recovery-act- 34 U. Energy Information Administration. U. National Science Foundation Energy Review 2010 (Washington. 37 North Carolina Solar Center and the Interstate Renewable Energy Council (IREC). Lauby et al.S. 2011). Department of Commerce. Department of Energy. Energy Review 2009 (Washington. 36 U. DC: U. “Workforce Training 47 2005). see note 25 above.. 2010). Department of Energy. (Washington.S. 2011). 2008). 1981–2002 Collaborative. Workshop on the Future Power Engineering Department of Energy. Annual Energy Workforce (Arlington.cfm. Data Needs for Public Policy Analysis (Washington.S. Supplemental Measures of Material Well-Being: Basic Needs. Annual workforce-training.gov/oe/ p23-202. Annual and M. 48 Department of Energy. Reder.S. et al. 2011).” http://energy. Energy. http://www. see note 41 above. U. 2004). Energy Information Administration.eia. Department Foundation. Power and Energy Engineering Workforce 46 Durables.S. Annual Energy Outlook 2011 (Washington.pdf.S.S. http//:www. VA: National Science Outlook 2011 (Washington. Energy Information Administration.S. DC: U.S. 30 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . n 32 U. DC: U. see note 40 above.S. Energy Information Administration. DC: U. Consumer U. 45 33 U.

Section 2. grid operators. but mechanisms for sharing data are immature. like wind and solar generation.Chapter 2: Enhancing the Transmission Network and System Operations In this chapter.1 provides an introduction to the transmission network and system operations. Sophisticated rate growing penetrations of variable energy new monitoring systems may reduce the resources. It supplies energy at high technologies may help solve problems associ- voltages to substations. The section also discusses promising emerging technologies. We find that the development of control algorithms that can utilize data from PMUs and exploit the capabilities of FACTS technologies are important areas for research. and improve system operations. This is followed in Section 2. More efficient or lower-impact electricity customers. It then explains how power systems are operated and briefly discusses transmission system reliability.2 by a description of technologies that could reduce the frequency of major blackouts. Section 2. sure to limit environmental impact. and many tools for data analysis have yet to be developed. Even so. this chapter does not deal with regulatory topics.3 introduces technologies that can facilitate the expansion of the transmission network and describes the fundamental physical characteristics that impose limits on transmission capacity and how these limits determine the technologies most appropriate for long-distance transmission. And but a variety of technologies offers the potential changes in system operation will help incorpo- to improve system performance. The transmission meet growing demand amid increasing pres- network today operates reliably and efficiently. including phasor measurement units (PMUs). likelihood of rare cascading system failures. then discuss new technologies that could help prevent blackouts. that can increase transmission network utilization and capacity. increase transmission capacity. we provide an overview of today’s transmission network technologies and power system operations. Section 2. The issues in this chapter are primarily the concern of utility engineers. and transmission planners. such as superconductors and dynamic line rating systems. including distributed to loads at lower voltages via the difficulty in siting new transmission lines to distribution network. accordingly. where the energy is ated with network expansion. we believe an understanding of transmission technologies and the operation of the power system provides important context for policy makers. and flexible alternating current transmission systems (FACTS). which are addressed in Chapter 4. The transmission network is the first link which can have serious economic and social between large power generation facilities and consequences. wide-area measurement systems (WAMS). Chapter 2: Enhancing the Transmission Network and System Operations 31 .4 describes a range of new technologies that could enhance system operations. or contain policy recommendations. We find that phasor measurement units have the potential to greatly benefit the transmission network. starting with transmission lines and substations.

With some exceptions. making it impossible to change the flow on one line without affecting others.000 by controlling miles of lines at 200 kilovolts (kV) and higher.2). not power. to transmission owners with only a handful of short transmission links. http://www.500 capacity—that is.S.1 Approximate U. see Box B.S. the transmission network is divided which energy flows is measured as power.800 the least-cost set of generators from supplying ≥ 600 0 load. linking electricity consumers to almost 5. power among lines lines cannot be sion network consists of approximately 170.000 to another follows multiple paths and may 300–399 54. ii Though it is energy. the U. Total ac 161. energy flowing from one location Alternating Current (ac) 200–299 84. transmis. Only weak electrical links exist the distribution of Power on individual between them.nerc. which control thou. how much power it can carry— by several mechanisms. Multiple Table 2. 1 Table 2.ii into three distinct geographic regions called Power is proportional to the product of current interconnections: the Western Interconnection. states.S.com/page. the rate at In the U. 32 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . we use the common industry vernacular in this chapter.1 contains a further explanation on the nature of power breakdown of transmission lines in the U. voltages at the two precisely controlled.1 in Appendix B). discussed in Section 2.. and voltage. Generators and other Reliability Council of Texas (ERCOT) (see devices manipulate Figure 1. higher voltage and current corre- the Eastern Interconnection.000 loop flows can create adverse or beneficial physical and economic effects in several ≥ 600 2. 2.400 jurisdictions. As a Line Type Voltage (kV) Miles consequence.000 ends of lines (for large power plants. Power on by miles and voltage level. These so-called 400–599 26. Altogether.3.php?cid=4|38.1 THE TRANSMISSION NETWORK AND Transmission Lines SYSTEM OPERATIONS Transmission lines carry energy. and the Electric spond to higher power. chapter are improving the ability of system sands of miles of lines spread over multiple operators to do so. Transmission Line Miles transmission lines often intersect at one by Voltage substation. flows. The interconnectedness of the grid compounds the difficulty in controlling power.i.S. own and operate these lines range from large though new devices discussed later in this investor-owned utilities.000 It is impossible to use the least-cost set of Source: North American Electric Reliability Corporation (NERC) Electricity Supply & Demand Database. that “flows” through transmission lines.000 cross jurisdictional boundaries. i Almost 5.000 Direct Current (dc) 200–299 700 The related problem of congestion results in 300–399 0 adverse economic consequences by preventing 400–599 1. it is common industry practice to speak of power flows rather than energy flows.000 generating units with at least 50 megawatts of expected on-peak summer capacity were registered with the North American Electric Reliability Corporation in 2010. A transmission line is limited in its Total dc 2. Total 169. The companies that individual lines cannot be precisely controlled.

and systems using transformers before reaching internationally. which is present in an ac system when current and voltage are not in phase. such as capacitor banks. provide more precise control and faster tioning of the transmission network and system response to changes in load. These lines may all be at the same voltage. which are discussed later in this connect transmission networks of different chapter.3 The decision to use an SVC customers. Finally. known as static VAR compensators (SVCs) sion substations is to interconnect transmission and are one member of a class of new devices lines. As a line is loaded—that is. economically justifiable. voltages. circuit breakers. common practice was to provide of total electricity billings in 2008. substations contain increased cost and energy losses of the SVC. could improve the performance of the transmission system.generators to supply additional load when one the voltage drop along the line from the or more transmission lines reach a limit and are generator to the load will increase. Maintaining voltage within a specified range along the entire length of an ac line may require special devices and control procedures. Interconnection was estimated to be about 6% Until recently. These devices are operations. of bringing the voltage back within acceptable When lines are thus congested. less range is known as voltage support or volt- economically efficient generators are dispatched ampere reactive (VAR) support. which if found to be acceptable limits. as its current is increased— iii VAR is the unit used to measure reactive power. measurement and communication equipment that bring data to control centers and voltage FINDING compensation devices that keep voltages within Technologies exist.2 voltage support by connecting compensating devices. other. comprising flexible ac transmission systems or the substation may contain transformers to (FACTS). The primary function of transmis. The first were typically decreased in several steps at substations installed in the early 1980s. substations provide instead of a capacitor bank is an economic one.S. Newer technolo- Transmission substations house much of the gies employ semiconductor switches and can equipment necessary for the normal func. the cost of congestion in the PJM as the length and loading of lines increases. to the line Substations and Voltage Support and controlling their voltage contribution in response to changes in load. protection for lines and equipment with devices the value of the operational benefits of fast and such as protective relays. Such costs can be significant in voltage levels and power transmission capacity some cases. These transformers also are necessary to connect the transmission system to the SVCs are an established technology with many lower-voltage distribution system. Voltage is years of operational deployment. Chapter 2: Enhancing the Transmission Network and System Operations 33 .iii Voltage to supply the load avoiding the transmission support is necessary to maintain acceptable network limits. and many more along the transmission and distribution installations have followed in the U. and flexible response is balanced against the surge arresters. In addition. The process unable to carry the required additional power.

In phenomena occur in fractions of a second. a similar result is obtained through bidding in wholesale markets. from system operators. A summary of the various generation and s Analysis: Raw data reported to control centers transmission operations and planning functions are analyzed using computer tools that can organized by timescale is presented in Figure 2. and power levels. west Power Pool. and fuel costs for generators. economic dispatch and unit commit. Remote terminal units also may receive of which generators should be on or off for commands. that planners at utilities and system operators generally address through appropriate mid- Control centers perform three separate and long-term planning. Transmission and generation owners that operate their own assets In addition to these functions. South. has the potential to give operators ment are calculated based on known start-up insight into these faster dynamics. Further discussion of transmis- s Monitoring: System operators use various sion planning policy issues can be found in displays and alarms to develop awareness Chapter 4. this chapter functions: focuses on transmission operations rather than on planning. Tennessee Valley Authority. PJM.3. discussed in utilities. This suite of tools is collectively by supervisory control and data acquisition known as an energy management system. including short-term and transmission constraints. System Operation areas. The The specification of the amount of power typical response time for SCADA systems today each of those committed generators should is several seconds. This Operator (ISO). called remote terminal units (RTU) located at ating units for the next day or days based on generators and substations collect this informa- projected electricity demand and relay this tion and send it to the control center every few information to generating units. (SCADA) systems that report the status of circuit breakers—open or closed—as well as s Control: Regional control centers calculate voltage. give insight to the current and future state of Currently. and control. the long-term and coordinate with these regional entities also health of the system is a separate concern are called system operators. For example. in restructured 34 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . ISO New England. The decision seconds. Midwest control mechanism is called automatic ISO. A single also give certain generators a signal that electrical interconnect contains many system supplements primary generator controls and operators. dispatching each reliability. However.1. System operators at control centers generator to minimize total system costs carry out many of these centralized functions given the load level. Section 2. generator availability. in support of operations. in the Eastern Inter. current. areas with traditional vertically integrated Important emerging technology. Devices the expected hourly power output of gener. close a breaker. Control centers Power systems require a level of centralized then refine these day-ahead estimates as often planning and operation to ensure system as every 5–15 minutes. but some power system produce is known as economic dispatch. such as an instruction to open or the next day is known as unit commitment. generation control. of the state of the system. and others. analysis. system control centers are supported the grid. enables the system to match small changes connection. system operators at the regional in load and meet the scheduled power level include the New York Independent System exchanges with neighboring systems. Control centers monitoring.

FL: CRC Press. Available data is environmental impact. and M. iv For example.S. sion network. Perez-Arriaga. not long enough to perform a good Research Institute (EPRI) has performed an evaluation. the Electric Power or less. “Electric Energy Systems: An Overview.000 megawatts in ERCOT must be reported. J. for example. assessing trends in the reliability of the U. increasing transmission capacity with low it is difficult to measure. and other stakeholders have put together a FINDING roadmap of innovative transmission technolo- Comprehensive and accurate data for gies. A. any loss of generation greater than 2. such as circuit breakers.S. but as discussed in Chapter 1.000 megawatts in the Eastern or Western Interconnection and 1. Gomez-Exposito. 2008). such as wind and solar power. operations.6 This figure is meant as a reference for the set of technologies discussed in this chapter. not a comprehensive list of transmission transmission network are not available. Note: AGC = automatic generation control. energy resources.” in Electric Energy Systems: Analysis and Operation. The most recent and comprehensive analysis of costs and benefits of various tech- report on these positive efforts is the 2011 Risk nologies5 and in Europe. the reli- ability of the U. are shown in begun to improve its practices for gathering black. are shown in red. Conejo. transmission grid clearly Reliability is and will continue to be a domi. Canizares (Boca Raton. Existing technologies NERC-defined “major events. Such lists of new subsequent legislation. 60. Transmission Reliability Although it is difficult to measure. A. much of the technologies have been compiled and discussed new data has been gathered for just a few years elsewhere. and C. equipment manufacturers.S. However. eds. transmission opera- Assessment of Reliability Performance. transmis. Important new technologies and func- and reporting reliability data in the last decade tions.Figure 2. universities.S. especially to incorporate variable term trends in reliability of the U. regardless of whether the loss of generation affects customers. and state estimation. Rudnick.” but these events and functions. address these challenges and shows how they fit Department of Energy (DOE) gather data on into the power system.iv NERC has formers. discussed in the remainder of this in response to the August 2003 blackout and chapter.2 lists technologies and functions to Reliability Corporation (NERC) and the U. faces challenges along three dimensions: nant constraint in transmission planning and anticipating and preventing blackouts. trans- do not necessarily affect customers.1 Transmission Operation and Planning Functions by Timescale Generator Primary Control Economic Unit Mid-Term Expansion Protection and AGC Dispatch Commitment Planning Planning Milliseconds Seconds Minutes Hours Days Weeks Years Source: I.4 tors. and improving system insufficient to make conclusions about long. technologies and functions. The North American Electric Figure 2. operations. Chapter 2: Enhancing the Transmission Network and System Operations 35 . Rivier. H.

POJUPSJOH"MBSNT t 4VQFSDPOEVDUJOH-JOFT t *NQSPWFE4ZTUFN t 'BVMU$VSSFOU-JNJUFST . enabling system operators to better protect against the most While major blackouts occur only rarely in the U. WAMS consist they have serious economic and social consequences. they have serious economic and social the state of the power system (see Box 2. Figure 2. communications networks.2 PREVENTING BLACKOUTS become unstable and subsequently affect a much wider area.FBTVSFNFOU6OJUT t 6OJU$PNNJUNFOU t 4FOTPST t $POUJOHFODZ"OBMZTJT -JOFTBOE$BCMFT t "VUPNBUJD(FOFSBUPS $POUSPM t &YUSBIJHIWPMUBHF"$-JOFT t )JHIWPMUBHF%$-JOFT t 1IBTF .S. and visualization software. The largest blackout in North Wide-area measurement systems (WAMS) American history occurred in 2003.2 Transmission Network Technologies and Control Center Functions Generators and Substations: t 1SPUFDUJWF3FMBZT t $JSDVJU#SFBLFST $POUSPM$FOUFS t 5SBOTGPSNFST t 4$"%"3FNPUF5FSNJOBM6OJUT t 4VSHF"SSFTUFST &OFSHZ. of measurement devices.BOBHFNFOU t 1IBTF4IJGUFST 4UTUFN'VODUJPOT t 'MFYJCMF"$5SBOTNJTTJPO t 4UBUF&TUJNBUJPO 4ZTUFNT t &DPOPNJD%JTQBUDI t 1IBTPS.S. Both blackouts phasor measurement unit (PMU). consequences.. critical is an enabling technology called the affected 30 million customers. were the result of cascading failures of the power system. catastrophic class of blackouts. 2. The second largest. in which seemingly small and localized problems caused the system to 36 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . Preventing such blackouts While major blackouts occur only rarely in the is an important goal that requires monitoring U. the most states and Ontario.1). new and emerging elements are shown in red. SCADA = Supervisory control and data acquisition.. affecting allow such monitoring to occur on a larger 50 million customers in eight northeastern scale than previously possible.PEFMT t %ZOBNJD-JOF3BUJOH4ZTUFNT t 0TDJMMBUJPO%FUFDUJPO Note: Existing technologies and functions are listed in black. in 1965.

limited by the small number of accompanying System frequency and other quantities often PMUs and software programs to process the also are measured. the PMU data and produce actionable informa- Industry standards require that PMUs have tion for system operation or planning are a reporting rate of 30 times per second. generators. Software applications to aggregate and analyze resolution information about system dynamics. The state these issues by exploiting the redundancy of of a power system is a snapshot of the system measurements throughout the system. One such proposed tool is a monitor or synchronized using GPS time signals. such as cities.1 in Appendix B. These data data acquisition system are not sent at the same are used in models of power systems that instant. several attempts to converge on a solution. As of early 2010. and critical to realizing the full benefits of PMUs. but their usefulness has historically been generators. voltages and currents at key substations. have trouble estimating a system state during unusual or emergency conditions—unfortu- nately. operators the estimated system state is obtained after may use the model results to identify anoma. and loads. they remain relatively undeveloped Critically. and power on all the system state. the transmission network stray beyond predict- able ranges. State estimators address output is the estimated system state. which results in higher.PMUs measure defining characteristics of WAMS are currently deployed in many areas. voltages and currents at one time that operators use to assess the condition of the system and. The lous system conditions. Chapter 2: Enhancing the Transmission Network and System Operations 37 . and thus phase angle data models are known as state estimators. For example. These several seconds. Another is that these data are not always precise. One reason is that sensor mea- lines in transmission systems are under continu. algorithm is not perfect. approximately line characteristics. BOX 2. and their cannot be observed. dispatch generation. The concept of phase angle difference is discussed in more detail in Box B.1 ESTIMATING THE STATE As with any model. data may be spread over a period of include the lines. measurements from all PMUs can be today. with more than 850 additional PMUs throughout the system. State estimators use an iterative algorithm. take action. many devices are capable of even higher rates. scheduled to be added between 2010 and 2013 through projects funded by the American PMUs report data much more frequently than Recovery and Reinvestment Act of 2009. Taken together with known raw PMU data. and state estimators and avoid stability and thermal limits. and if needed. However.7 do SCADA systems. currents. surements from the supervisory control and ous monitoring by system operators. the result of the state OF THE POWER SYSTEM estimator is only an approximation of the actual The voltages. indicating that the system is under stress. these measurements can be 250 PMUs were deployed across North used to calculate instantaneous power flows America. when they are most needed. enabling alarm that would warn when the voltage phase more accurate characterization of system-wide angle differences between different locations on dynamics. and load centers.

NERC created two an opportunity to take remedial action. One agreement is meant for industry entities The development of phase angle alarms is not and covers the confidential sharing of phasor trivial. NJ. system operators would have had To facilitate sharing of synchronized phasor more warning of the impending problem and data between regions. Obtaining between Cleveland and Michigan leading up to baseline data through observation is an impor- the 2003 blackout. 2010). Analysis of phase angle tant prerequisite for not only phase angle measurements revealed a slow divergence alarms but also for many other potential nearly an hour before the start of the blackout. for as a whole.com/filez/rapirtf.v phase angle monitoring application been in use at the time.8 nondisclosure agreements in February 2010. software applications of synchronized phasor Had the PMUs been networked and a real-time measurements. Only a few early phase restricted basis for the benefit of the industry angle alarms have been implemented. Real-Time Application of Synchrophasors for Improving Reliability (Princeton.9 It Can Fall Apart in Three Minutes 38 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . v One other specific benefit of synchronized phasor measurements is mentioned in the next section on increasing transmission capacity. example. The usefulness of phase angle alarm applica.html.3 Cleveland–Michigan Phase Angle Difference Leading Up to the August 2003 Blackout 140 120 100 80 Degrees 60 40 Cleveland – Michigan Angle 20 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 25 26 27 28 29 30 31 32 33 34 35 36 37 38 38 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 Minutes Source: North American Electric Reliability Corporation Real-Time Application of PMUs to Improve Reliability Task Force. Baseline data must be collected from data within the NERC phasor community PMUs over a minimum of several years before for operational and reliability purposes. http://www. the Bonneville Power Administration Figure 2.3. has had a system in operation long enough to tions is illustrated in Figure 2. which shows establish a baseline for the phase angles at three the phase angle difference as a function of time of its hydroelectric generators. alarms or other useful applications can become The second agreement covers the sharing operational.nerc. and these data must be shared with of industry phasor data with researchers on a all relevant stakeholders.

and transient network. As of September 2011. the transmission network’s capacity must be enhanced to reliably and of a transmission line: the thermal constraint. obtain for political and environmental policy To determine these stability limits. It is clear that for any And even planning new transmission lines— set of normal system conditions. Projects can stretch over many years. only strategic regional and interregional objectives. policy makers building new lines. power flows a subject covered in more detail in Chapter 4— must not cause overheating or system instability. 2.2. must perform an extensive analysis known as an delaying needed network reinforcements. Transmission Capacity Limits FINDING Phasor measurement units have the There are three primary constraints on the capacity of a transmission line: the thermal potential to greatly benefit the transmission constraint. and superconductors. a limited number of entities had signed these data-sharing agreements. but in order to effectively use transmission capacity. New rights-of-way—the land on reached when operators are concerned that an which lines are built—are very difficult to unexpected event might cause system instability. high-voltage transmission overlays. operators reasons. tool applied to this task is the construction of new transmission lines. The first. These include extra-high-voltage (EHV) ac and high-voltage direct current (dc) These concerns do not warrant immediate lines. and many tools for data straightforward. we present trans- the data. information about the underlying mission line technologies that can increase network is also required. and sufficient sharing of network models Finally. mechanisms for sharing stability. As load centers grow and generators are built There are three primary constraints on the capacity in new locations. The losses in a line increase its analysis have yet to be developed. the thermal constraint. or thermal constraint. voltage economically connect the two.3 INCREASING TRANSMISSION the sag is sufficient to reduce the line’s clearance CAPACITY from ground to a minimum acceptable value. is data are immature. new. Stability limits are more complicated than the mental building block of the transmission thermal constraint and derive from consider- network. bottleneck. action because the agreement is still relatively underground cables. Building new transmission lines can ations discussed in Box 2. capacity. The primary stability. voltage stability. At some maximum temperature. N-1 contingency analysis. though these are not should be cognizant of this potential research long-term substitutes for new infrastructure. we discuss the potential of PMUs and may occur organically over time between dynamic line rating (DLR) systems to increase system operators and some trusted academic transmission capacity in some situations without research institutions. for they are the funda. we first discuss the fundamental is that the agreements cover only sharing of physical characteristics that impose limits to measured data.It remains to be seen whether these agreements of local transmission networks with long-term will prove effective. temperature. However.10 A further concern In this section. and transient stability. These limits are be difficult. is complex and requires balancing the details N-1 contingency analysis takes this one step Chapter 2: Enhancing the Transmission Network and System Operations 39 . Next. which in turn causes the line to stretch and sag between supports. However.

Instability can lead to major negative across the network after a disturbance. Due to the large number of possible variable energy resources. transmission a disturbance when the automatic controls capacity. and tightly meshed networks contrib. transmission line to accept a transient increase further to ensure that even if any single major Long-Distance Transmission Technologies system component (such as a large generator or transmission line) is unexpectedly lost. constrained by voltage stability. is a progressive drop in voltages following adequate reserve generation. discussed in Chapter 4. and able simplification of the precise industry stability problems of one sort usually give rise standard classifications. A comprehensive the physical nature of stability phenomena.. regulation. into the grid. The consequences. and the methods of analysis that Stability IEEE/CIGRÉ Joint Task Force on Stability must be used to address the stability issues. Thermal considerations generally limit discuss the two technologies appropriate for power transfer on short lines. and transmission equipment beyond their Two of the main forms of stability that concern capabilities. while the longest lines are limited by transient stability. Here we tions. The thermal limit has to do with balanced with the higher cost and difficulty of material properties of the line and is constant siting such lines. two relatively transfer on medium-length lines is usually mature technologies. These energy resources have occur. Since stability transmission for distant renewable generation is a system property rather than a material sources is a difficult problem to be addressed property. Stability classifications are based on other types of stability. they are often distant from load centers. gained favor at the federal and state levels. Striking the proper balance in no matter the length of a line. the Challenging these physical constraints is power flows on network lines still do not violate increasing political interest in integrating these limits. from localized power interrup. 40 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . this analysis requires expert judgment. Terms and Definitions. associated with some loads push generators ute to a stable system. Kundur et al. and computational power.” IEEE Transactions on Transient stability refers to the ability of a Power Systems 19 (2004): 1387–1401. The technical capability exists to provide long- Figure 2. system to maintain acceptable voltage levels bance. most common form of voltage instability tions to widespread blackouts. but the benefit of factors on the power transfer of short. system operators and planners are transient and The description of stability here is a consider- voltage stability. BOX 2. medium. In general. They are interrelated.2 POWER SYSTEM STABILITY in power flow without exceeding the maximum Stability of an alternating current power system safe voltage angle between the ends of the line. the description can be found in P. refers to its ability to maintain synchronous Voltage stability refers to the ability of a power operation after being subjected to a distur. In particular. such as wind and situations under which contingencies might solar. which include several to others.4 illustrates the typical limiting distance transmission. stability limits change depending on through transmission planning policy and the length of a line and other system condi. but time. power voltage ac and high-voltage dc. system variables where that stability phenomena “Definition and Classification of Power System are observed. while longer lines long-distance transmission—extra-high- tend to be stability limited. access to distant renewable resources must be and long lines.

As a limited by stability considerations and therefore result.11 Higher voltage trans. 12 due to their many desirable networks. the highest-voltage ac lines are most theoretically are not limited in length. economical for large-capacity. Dc lines can be valuable mission lines have been installed in China.S. The length and capacity of a dc substations is significantly higher because long EHV ac line is typically limited by stability transformers only work for ac. and Japan. is 765 kV. and sion networks. High-voltage dc lines are not towers.4 Three Primary Constraints of Transmission Lines Length of Line Limits: Thermal Constraint Long (More than Voltage Stability 150 miles) Transient Stability Medium (50–150 miles) Short (0–50 miles) Limited Limited Limited by transient by voltage by thermal Maximum Power stability stability constraint (overheating) EHV ac transmission systems have voltages considerations—the longer the line. Installing high-voltage. while 345 kV and 500 kV are support equipment.000 kV system at its rated voltage.vi. However. But the cost for of the network. conducts electricity—for dc transmission lines large-capacity links can improve reliability by are lower than for ac lines of the same voltage allowing neighboring areas to support one because fewer conductors are necessary and another and improving stability characteristics conductor utilization is better. though the effective length The highest in commercial operation in the of lines may be extended by installing voltage U.Figure 2.000 kV. so more vi Russia and Japan now operate their 1. The transmission system Russia. insulators. long-distance Conductor costs—the cost of the metal which electricity transmission. Chapter 2: Enhancing the Transmission Network and System Operations 41 . the capacity limit. dc power over longer distances but require larger. as well as wider rights-of-way. the lower greater than 242 but less than 1. lines can be valuable additions to ac transmis- more expensive transformers. but only China operates its consists mainly of ac lines additions to ac transmission 1. Compared to their lower-voltage counterparts. standard voltage levels.000 kV lines at 550 kV. characteristics. such as the such lines are capable of transmitting more ease of voltage transformation.

known as voltage source converters (VSC) offer Thus. the capacity of electrical losses of the 765 kV ac line at full load this link is only 300 MW. 42 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . Thyristor switches can be turned on easily but may only turn off under certain conditions. while combined lines may be economically attractive. least-expensive option is not necessarily the most appropriate. the losses per mile A newer version of high-voltage dc converters of a dc line are lower than those of an ac line.vii A dc mesh network using 1. FINDING A rough rule of thumb is that for high-voltage Where long distances separate renewable ac lines.000 mega- dc: the highest-voltage link in operation as of watt transmission link at different distances. expensive power electronics converter stations high-voltage ac and dc transmission projects. Transistor switches can be turned on and off easily. making it more attractive as a technology to connect In the U. however.16 And unlike conventional high-voltage single generating site connects to a single point dc. though the efficiency and have approximately the same losses.5%. the ±500 kV Pacific DC Intertie variable energy resources. This of VSC converters is also improving. China’s ±800 kV dc link from that is used in the converter—a transistor rather the Xiangjiaba Dam to Shanghai is nearly than a thyristor.15 2011 is the ±350 kV Caprivi Link connecting For an 800-mile link. VSCs do not require strong ac generation on the ac grid.300 miles long and currently the world’s VSC high-voltage dc is envisioned to connect longest and highest-voltage dc link. but they have different characteristics. are required to convert between ac and dc..18 VSC double the cost of the ±800 kV dc line. but new advances in would be nearly double those of a ±800 kV semiconductor devices and dc circuit breakers dc line and that the up-front cost to build the promise to allow higher voltage levels and 765 kV ac line would also be approximately capacities within the next few years. the potential for improved system stability and point-to-point power transmission. sources at both ends of the line. this One analysis of hypothetical transmission technology has not achieved the highest projects compared the cost and electrical losses voltages attainable by traditional high-voltage of 765 kV ac to ±800 kV dc for a 6. the analysis found that Zambezi and Gerus in Namibia. dc is especially suited to long-distance. If the high-voltage dc converter stations have some- project were a line of only 200 miles. such as long-term system impact on complex converter station instead of a much reliability and right-of-way considerations.17 So far. case is merely illustrative of the relationships between length of lines and cost/losses of vii The transistor and thyristor are both semiconductor devices that function as a switch.5% of rated power. And any real project would be evaluated in a more tapping a dc line—that is. The electrical losses in an ac/dc converter station are higher than in an ac substation. total substation losses are approxi- resources from load centers. These benefits are stretches nearly 850 miles from Oregon to made possible by a more flexible type of switch Los Angeles.S. dc transmission mately 0. where a control. The less expensive transformer as for an ac line. connecting a load detailed fashion on many more important in the middle of the line—requires a costly and criteria.14 dispersed wind generators in the North Sea with several areas of mainland Europe. what higher energy loss than conventional the dc option would be slightly more expensive high-voltage dc stations. converter station losses from both ends of a traditional high-voltage dc line are approxi- mately 1.13 However.

improved reliability. tightly meshed network spanning a large this study concluded the optimal solution to be geographic area to facilitate the integration a transmission overlay serving wind zones of of variable energy resources.”20 For a given level of power of renewables is technically feasible through transmission.S. These standards dictate that a certain percentage of energy is generated from renewable resources. including access to better As noted in Chapter 4. the both types. and optimally loaded lines have lines. rather than only one or the other: benefits of such an overlay. the features and costs of an optimal described in Chapter 4. described here. some local wind resources and some distant. proposals have been made for resources local to load centers and favorable high-voltage transmission overlays—a new capacity factors of wind resources distant from network of EHV transmission lines superim- load centers. It is also possible that stronger ties between areas would FINDING allow system operators to reduce requirements for costly reserves. Interregional renew- System Outage Task Force notes.19 Figure 2. A transmission overlay undoubtedly would have many benefits. per delivered megawatt hour. losses decrease when voltage is higher-voltage. planning processes sources of renewable energy. This trade-off is illustrated in posed on the existing transmission network. Chapter 2: Enhancing the Transmission Network and System Operations 43 . such as the Eastern lines and more tightly networked lines…are Wind Integration and Transmission Study and better able to absorb voltage and current swings the Western Wind and Solar Integration Study and thus serve as a barrier to the spread of a have shown that integrating high penetrations [cascading failure]. the study went on constructing such an extensive transmission to analyze transmission overlay options and network. but a true plan has yet to emerge. However. for interregional areas face both technical and and lower losses.5. Through detailed analysis of the Such transmission overlays could create a more particular characteristics of the Midwest region. must be valued against the very high cost of Based on this conclusion. projects to be used as inputs in the Midwest ISO’s transmission planning process. Evaluating the full range of costs the interconnection of renewables.FINDING example of a successful system planning The control flexibility of voltage source exercise at the regional level.22 lower losses than overloaded lines.–Canada Power institutional challenges.21 The purpose converters can improve system stability and of the study was to support development of facilitate the integration of remotely located transmission portfolios fulfilling the region’s renewable portfolio standards at the lowest cost renewable generation. Creating a high-voltage transmission overlay is technically feasible and would But building a network of high-voltage lines benefit system operations and facilitate is also costly. In the and benefits of a transmission overlay requires absence of detailed inter-regional planning the sort of interregional planning process studies. The Midwest ISO’s Regional Generation Outlet Study is an future overlay network remain uncertain. a cost that we do not presume identify a set of promising transmission to estimate. As the U. One key Transmission Overlays consideration noted in this study is the balance between low transmission costs for wind In recent years. tightly meshed transmission increased. “higher voltage ables integration studies.

This project connected two substations sepa- rated by 600 meters in the Long Island Power 44 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . 2010).5 Trade-off between Transmission Cost and Capacity Factors High Capacity Cost High Transmision Cost Low Transmision Cost Low Capacity Cost Total Minimum Total Cost: Cost ($) Energy. a relatively inexpensive coolant kilometer link between Norway and the with a boiling point of -196°C. was successfully demonstrated in an attractive option in some areas despite the 2008 with the help of government funding. Figure 2. Underground and Submarine Transmission Superconductors Underground or submarine cables are used in High-temperature superconductors (HTSCs) locations where overhead lines are impossible have emerged from the research labs within the or undesirable. They have a Netherlands.S. generally limits their length to just tens of which is different for each superconductor. IN. miles. the complexity of assembling and normal conductors of the same physical size. the longest submarine dc cable is a 580 liquid nitrogen. installing cables means that cables will remain but are constrained by the difficulty of main- more expensive than overhead lines. Dc cables are limited only by electrical HTSCs are those that may be cooled using losses. Despite innovations in insulation much higher power capacity compared to materials. A severe constraint when these past decade. highest- the difficulty of siting overhead lines in the U. Capacity. The longest.24 greater expense. operating at can make underground and submarine cables 138 kV ac. and Transmission H Capacity Cost L L Transmission Cost H Source: Midwest Independent System Operator.23 However. capacity HTSC cable to date. Regional Generation Outlet Study (Carmel. Superconductors are materials that cables are used for ac transmission is that the have extremely low electrical resistance when high capacitive charging current required cooled below a certain critical temperature. taining adequate cooling.

Chapter 2: Enhancing the Transmission Network and System Operations 45 . This static limit is two years. In the case of lines could result in large economic benefits. the purpose of an FCL is to a few key transmission lines could reduce the amount of current that flows under between the California fault conditions—not by isolating it through ISO and NERC to share result in large economic benefits. the impedance is inserted by exploiting the natural Dynamic Line Ratings limits of superconducting materials. switches but by introducing a high impedance synchronized phasor to reduce the level of current. the material Dynamic line rating systems also potentially reverts from superconducting behavior to a can increase the operational capacity of trans- normal conductor having high impedance. In addition to helping prevent blackouts. However. the grid closer to its true stability limit.26 often conservative relative to actual conditions. Since high impedance is not desirable expected to be common because only stability- under normal system conditions. On some lines.25 High- temperature superconductors are becoming Phasor Measurements for Increased more practical as an option for increasing the Transmission Capacity capacity of existing cables by replacing them with ac HTSC cables in the same conduits. increases in in the system extremely quickly only when it is the capacity of only a few key transmission needed. system operators While promising. HTSC cable technology has been conducting transmission lines. the refrigeration system and the ability to make joints between cable sections. one simple superconducting FCL design. under fault conditions. DLR systems measure changing environmental viii This is also true of FCLs constructed by other means. A second stage of this project is under way that will test rapid field repair of without the need for new rights-of-way. In an early Increases in the capacity of only example. effec- tively increasing capacity of some lines without Superconductors also may be used to create a increasing the risk of a device called a fault current limiter (FCL). an agreement its name implies. mission lines. Above some maximum current level. This is particularly attractive where the cost synchronized PMUs can potentially improve of conventional alternatives is high. As blackout. data is expected to eventually result in a 30% adding an FCL can allow heavier line loading increase in the capacity of the California– without exceeding the capacity of circuit Oregon Intertie. such as in system capacity by allowing operators to take dense urban areas. the challenge limited lines have the chance to benefit from in designing an FCL is how to insert impedance the synchronized data. under seasonal worst-case assumptions. superconductor FCLs have have established the thermal limits of lines not yet been demonstrated at transmission. Historically. a hot. such as using power electronic devices.Authority service territory and overcame FINDING several practical technical challenges of super.27 Such large increases are not breakers. such as the demonstrated as practical and is a ability to withstand fault currents—abnormally promising approach to substantially high current levels caused by short circuits on a increasing the capacity of existing cables transmission line. level voltages.viii though at least two projects windless day is an example of a worst-case hope to achieve this goal within the next scenario in the summer.

resulting in more optimal transmission capacity than traditional static line ratings. the main determinant 2. However. increasing transmission capacity limits in in 2009 revealed that only 0. the 2003 blackout. where wind resources are far Energy Management Systems Integrating from load centers.28 demand response schemes (Chapter 7) will create opportunities for technological change Dynamic line rating systems installed on existing in system operations. when static transmission these improvements are incremental and do capacity limits for some lines had been set not address the fundamental shortcoming that assuming a modest amount of wind would cool state estimators sometimes fail when system wires.” 30 Based on the positive results of deployed relatively widely today uses just two. DOE dynamic line rating systems can be imple. previous deployments. which DLR systems improve transmission capacity.5% of respondent’s all but those few worst-case scenarios. when in fact there was hardly any wind. systems of the future are less understood than Wind generators require more transmission many individual technologies discussed in this capacity when the wind is strongest. when electric loads are high and systems are their most stressed. Though lines were equipped with DLR systems. Because a survey of electric service providers ingly. Notably. conditions and update system models accord. DLR to existing state estimators can improve accuracy systems will place more restrictive limits on of the estimated state. technologies. has classified the penetration of these systems mented with a variety of sensors.29 conditions are unusual. These two pieces of between now and 2030. chapter and should be the subject of R&D strong winds are precisely the conditions in efforts by utilities and academic institutions. Knowledge of true transmission The control systems of the future are capacity limits means improved reliability at less understood than many individual the most critical times of extreme weather. novel communications architectures. DLR systems installed on existing transmission lines have been shown The challenges of intermittent generation to improve the capacity by 5%–30% depending (Chapter 3) and opportunities for advanced on conditions. This was the case during economic dispatch of generating units. we can expect an one to measure line tension and another to increase in penetration of DLR systems measure air temperature. Adding synchronized phasor data In some extreme weather situations. The control tors to the rest of the transmission network. Eventually state estima- A DLR system would have provided another tors could be partially replaced by aPMU-based layer of warning to system operators. energy manage- ment systems will need upgrades to accommo- transmission lines have been shown to improve the date synchronized measurements from PMUs. conveniently.4 IMPROVING SYSTEM OPERATIONS of a line’s thermal limit. tool yet to be developed that directly measures 46 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . could play DLR systems are particularly attractive in the a role in advanced control schemes requiring case of transmission lines linking wind genera. data allow operators to determine average conductor temperature. However. And power electronics devices. DLR systems Energy management systems must be updated will not improve the capacity of these lines. capacity by 5%–30% depending on conditions. one design as “nascent. the long connecting lines PMUs needed would tend to be limited by stability rather than thermal properties. if they are to process the additional data available from PMUs. supported by these synchronized measurements.

On the other hand. Chapter 2: Enhancing the Transmission Network and System Operations 47 . Such responses can include inten- likely supplement rather than replace current tionally islanding the two regions. Among these are more sophisticated important. availability of more accurate and timely data control room visualization techniques and make possible new approaches to system operator training will become increasingly control. load. problem. mously complex. but present old and new schemes (SIPS). Additional tools in the control The most complex control schemes deployed room must not simply make new information today are called system integrity protection available to operators. quickly assimilate a staggering amount of information. would responses. FINDING Eventually. and would avoid regions. within which load and generation are balanced. faster and more accurately than currently wide-area control could be nothing more than employed state estimation tools.32 which are comprised of information in more effective ways. as a result. such rest of the system—SIPS use precalculated a system might be feasible only at the highest scenarios to coordinate more intelligent voltage levels by 2030 and. wide-area control could also use closed-loop feedback control to stabilize detected system oscillations. Determining where such islands might be created to aid system stability requires careful study of system contingencies.rather than estimates system state.31 mented some SIPS. system operators primarily This tool would require PMUs to be deployed in the Western Interconnection have imple- very widely. Advanced Control Schemes As these research efforts bear fruit and new Increased computational power and the tools become available to system operators. decentralized subsystems that make decisions operators must train extensively on new tools based on local and wide-area measurements. and creating such an island involves switching many circuit breakers and also likely shedding some amount of load within the island to balance load and generation. A more what SIPS are today: protective procedures extensive deployment of PMUs is required developed to respond to a specific type of to make this possible. Such a tool These system-wide SIPS are normally imple- would be faster and potentially more accurate mented in cases of large power transfers between than state estimators today. wide-area control systems. when exceeding that line’s capacity the problems state estimators have finding a rating could potentially trigger a catastrophic solution during unusual system conditions. new ix An island in this context is a self-contained section of the network.34 The energy management systems by providing concept encompasses a broad range of possible real-time data to determine system state future control schemes. operators must and closed-loop control using PMU data. or activating voltage support devices. and in a crisis. Where a typical protection Concerns about the reliability and accuracy of system would simply isolate the offending line PMU measurements for critical applications by tripping circuit breakers at either end— would also need to be addressed. at perhaps 30%–50% of all nodes. Additionally. blackout. On the one hand.S. Given the potentially sending shock waves through the number of PMUs being deployed today.33 before they become operational. The control room is already enor- protection actions. if developed.ix shedding state estimators. PMU-supported WAMS could PMUs could improve the performance of be transformed into wide-area control systems that actively participate in control actions. In the U..

Integration into more sophisti- in combination with detailed simulations of cated control systems could help justify these potential crises that might occur. summarized in Table 2. NERC standards require a more systematic transmission systems. Research on FACTS can be divided into three categories: FINDING semiconductor materials.35 Training includes general familiarization because of cost. in power systems of the future. research in control algorithms and which become obsolete with changes to the improved confidence in the reliability and transmission network. and time required to develop appropriate visualization tools and prepare operators to use R&D efforts to reduce costs will be necessary them should be incorporated into R&D strate. Research collaborations among system operators. FACTS devices. The resources high costs in some situations. the deployment approach to training than did previous stan. and other new the transmission network to enable more rapid technologies. of devices other than SVCs has been limited dards. However. Today.2 Summary of Main FACTS Devices Name Most Suitable Functions Static VAR Compensator Control voltage level at nodes Static Synchronous Compensator Improve system stability characteristics Thyristor-Controlled Series Compensator Improve system stability characteristics Control flows of power Unified Power Flow Controller Control voltage level at nodes Control flows of power Improve system stability characteristics 48 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . control systems leveraging and flexible control of the system. These benefits of synchronized phasor employ power electronics that are connected to data. Flexible Ac Transmission Systems FINDING A critical piece of the advanced control schemes To fully realize the improved system envisioned for the future is FACTS. such systems are limited Work is also needed to develop strategies to in number and capability. time data from a wide-area network of More work is needed in all three areas. operations.2. academic institutions. control algorithms Automatic control action based on real. within individual FACTS devices. Each of the devices listed in the table has been deployed on real Table 2. The basic the complementary features of these characteristics of several FACTS devices are technologies need to be developed. but the PMUs represents a major change in system last two categories are particularly important. Significant replace highly specific control algorithms. if FACTS are to become a significant factor gies for integrative control systems. with algorithms based accuracy of PMU data is needed to make on a reconfigurable architecture. and equipment vendors should be encouraged. such controls more prevalent. and system- wide control schemes incorporating FACTS.

acting alone. tion tasks on the power system today. gather information and make intelligent But new technologies are available that can operational decisions autonomously. recent work has explored the and after Hurricane Irene in August 2011.36 the transmission network using switches— along with generator production in power More disruptive changes could occur in the flow models. latency. Over the past decade. and privacy. electricity markets. Some researchers envision an IT framework that As noted in Chapter 1.38 New approaches to transmission and distribution systems across other optimization problems in system opera- multiple jurisdictions. this of multiple technologies into an integrated would require more seamless sharing of system of sensors. the combination greater knowledge. No electricity or other aspects of system status and single technology.S. the potential of leveraging modern communi. possibility of optimizing transmission a significant improvement over information switching—adding or removing lines from available during the August 2003 blackout. an electricity customer might have enhanced reliability. One visualization system has begun to show advantage of such gains. Among other things. Chapter 2: Enhancing the Transmission Network and System Operations 49 . These efforts have and draws on data reported by multiple utilities resulted in considerable cost savings in several to provide up-to-date information on both the U. To operations will be different groups of stakeholders accommodate the high bandwidth. offering example.39 architecture of the communication system. of computers and new make intelligent operational algorithms to take decisions autonomously.5 CONCLUSIONS AND place at the control center based on data RECOMMENDATIONS gathered through the SCADA system. the transmission allows different groups of stakeholders to network today operates reliably and efficiently. PJM has led an effort cations infrastructure to the direct benefit of to develop and implement new optimization electricity customers. and supported by gains in reliability needs of future software applications. communications x Chapter 9 is devoted to issues surrounding communications.37 For improve system performance. level of reliability. the processing power to gather information and fiber optics likely will become more prominent. The Energy Awareness algorithms to help solve the problem of how and Resilience Streaming Service (EARSS) was many generator units to commit each day. power line carrier. data security. However. support emergency and recovery efforts during For example. The big challenge for New communications infrastructures x and regulators in deciding whether operators architectures will support power system should cede some control to other stakeholders operations in the future. is likely to have optimize electricity consumption based on this a significant impact. increased capacity. radio. developed at Oak Ridge National Laboratory and when to commit them. most system operational decisions take 2. and fiber optics New tools for system an IT framework that allows are some of the more common media. Many methods of data is how to maintain and guarantee the same high transmittal are used for various communica.Information and Communication information between transmission and distri- bution system operators. Some researchers envision microwave. and access to information about the price of the ability to better accommodate VERs. Today. The system was used to tions might result in further substantial savings.

Confidential data sharing agreements have been PMUs integrated into wide-area measurement created by NERC. and intel. but only a limited number of systems with appropriate analysis tools that relevant entities have executed them. However. area measurement systems. essential to development of the needed tools. be engaged in the development of the technolo- sive deployment of FACTS devices. integrated with PMUs and wide-area measure- ment systems. and researchers is significant benefits. infrastructure. The inte- gration of FACTS devices with emerging wide-area measurement systems can allow their control capabilities to be leveraged to provide even greater benefits and could make their costs more readily justified. and is being installed relevant entities to participate in PMU data- more widely as a result of ARRA funding. their full benefit control capabilities of these devices can will occur only if there is greater cooperation contribute significantly to network control if among utilities in their deployment. the current high costs of the most versatile of these devices is inhib- iting their widespread deployment. While NERC should continue to encourage PMU hardware exists. turn the measured data into actionable infor- mation could provide protection from black. R E CO M M E N D AT I O N Research and development efforts should be undertaken to develop 1) the analysis tools necessary to generate actionable information from data acquired from PMUs and 2) the control schemes necessary to make use of this information by realizing the complementary potential of PMUs. The rapid gies discussed in this chapter. R E CO M M E N D AT I O N outs and increases in system capacity. control equipment. 50 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . Greater control of system voltages and power Although industry has been and continues to flow can be achieved through the more exten. the sharing efforts necessary for the effective software and analysis tools necessary to fully capitalize on this investment are yet to be development and use of PMUs and wide- developed and deployed. and other hardware devices. system operators. More widespread PMU data sharing among ligent management systems will provide utilities. FACTS.

Liu. 2010. W. Åström. report http://www. CA. applications. National Electric Substations in Japan.abb. Johansson. Eastern Wind Integration and 19 Transmission Study. DC. 2009). Uehara. prepared for the National Renewable Energy 11 IEEE Standard Preferred Voltage Ratings for Laboratory (Golden. “EHV AC and HVDC 15 Report on Cost/Benefit Analysis of Innovative Transmission Working Together to Integrate Technologies and Grid Technologies Roadmap Report Renewable Power. Regional 21 technical report 1015511 (Palo Alto. E.2. 2011). Fleeman et al. Germany. Department of Energy. The ABB Group. Generation Outlet Study (Carmel. and R. 2007).REFERENCES “Focus on UHV AC: China Shows the Way by 12 Energising 1. see North American Electric TN. September 10–11. and T. 2011). July 19. Out of Mind 23 Equipment Operating at Voltages Above 230 kV Revisited: An Updated Study on the Undergrounding Nominal.com/industries/ap/db0003db004333/718 Recommendations (Washington.S. K.” Western Wind and Solar Integration Study. Estimating the U. of Overhead Power Lines (Washington. REALISEGRID Wide-Scale Renewable Resources into the Power Deliverable D1.” http://www. Sasamori. K. Hyttinen. page 75.” Global Transmission 1 North American Electric Reliability Corporation Report. A.com/page. and Electric Power Research Institute.4. 2003 Blackout in the ABB. Pagano. 4 North American Electric Reliability Corporation.nerc. see note 19 above. March 2. http:// globaltransmission. Chapter 2: Enhancing the Transmission Network and System Operations 51 . “HVDC Projects by Type and Power. 2010).S. http://www.S. IN.000 kV line. Y. and V. press release. Electricity Supply & Demand Database.com/page. Lescale. and Z. Ma. http://realisegrid.4. Reliability Task Force.php?cid=4|38.nerc. “Power System Stability Benefits with Real-Time Application of PMUs to Improve VSC-HVDC Transmission Systems” (Paris: CIGRÉ. “The ABCs of 13 3 Siemens.2 (June 6. 8 U. doi:10. Electrical Power (2011). M. Real-Time Application 2004). Bornholm. DC. note 7 U. P. DC: U.” IEEE Power List” (Erlangen. Final WP1 J. “Important Technologies Applied for UHV AC 2 U. Special Aspects” (Paris: CIGRÉ. Measurement Unit Implementation and Applications. Vaféas.. CIGRÉ/IEEE PES Joint rse-web.S.” European Transactions on Transmission Congestion Study (Washington. report prepared for the 9 For more detail on other envisioned PMU National Renewable Energy Laboratory (Knoxville. Yamagata. bfd4f5d7fa84bc12574ad00302100.” presented at the filez/rapirtf. 20results/Deliverable_REALISEGRID_1.html. Denmark. https://reports. Bahrman and B. Ono. Phasor see note 8 above. “ABB Commissions World’s 14 “Risk Assessment of Reliability Performance. M. IEEE Standard 1312-1993. Asplund. “The Xiangjiaba- Costs and Benefits of the Smart Grid. G. EnerNex.gov/BlackoutFinal-Web. technical Shanghai 800kV UHVDC Project Status and report 1022519 (Palo Alto. “Reactive Power Compensation Reference HVDC Transmission Technologies. Johnson. Reliability Corporation Real-Time Application of PMUs to Improve Reliability Task Force. Department of Energy. of Synchrophasors for Improving Reliability E. and www. 2009. i-2. EnerNex. 16 7 North American Electric Reliability Corporation Rudervall. http://www.php?cid=4|37.” http:// 18 United States and Canada: Causes and www.” presented at Integration of Validated by External Partners. G. S.com/page. Nordic Wind Power Conference. 2011). 2010). 2009. Out of Sight. CO. 10 North American Electric Reliability Corporation. CA. and K.–Canada Power System Outage Task Force. & Energy Magazine 5 (2007): 32–44. 2009). Alberta.aspx. Delivery System. 5 Electric Power Research Institute. 2010).nerc.–Canada Power System Outage Task Force. Jansson. DC. Koldby and M. 2009).info/archive. energy. 6 A.it/content/files/File/Publications%20and% Symposium. Calgary. Midwest Independent System Operator.nerc. and GE Energy. July 29–31. 2009.pdf.” Longest and Most Powerful Transmission Link.1002/etep. 2010).pdf.568. Galant. S. 2004).com/ Road to an Offshore HVDC Grid. Final Report on the August 14.php?cid=6|319|345.php?id=1434. “Challenges on the 17 (Washington. 22 “NASPI Data Sharing Agreement Signatories. 20 above. Alternating-Current Electrical Systems and Edison Electric Institute. F.

P. S. Begovic. Hedman. and K. 18. Xu and A. R. Island Power Authority (LIPA) Grid. “3Rs for Power and Demand. Bose..” presented at Power Systems Conference and 28 K.” presentation to CIGRÉ Symposium. S. Guido. “Flexible Transmission in the Smart Grid: Optimal see note 8 above. 2005). 35 North American Electric Reliability Corporation. “Progress and Status of a 2G Supporting Infrastructure. 31 B. Fault Current Limiters. Streiffert. Palmer “Grid Visualization Efforts Helped Heal After Hurricane. December 2009.” in Power 26 Electric Power Research Institute. 2007 iREP Symposium. 2005.” IEEE Spectrum. “System Integrity Protection Schemes (SIPS). Seattle. 32 A. Hur et al. and A. W. (Palo Alto.” presented at Bulk Power System Dynamics and Control—VII.ieee. August 19–24. 2007). W. S. P.. technical report 05-58 (Tempe. Maguire et al. 36 K. Croatia.–Canada Power System Outage Task Force.” IEEE Power Exposition. Oren. “High-Wire Act. Integer Programming Solution for Market Clearing and Reliability Analysis. Department of Energy. Abur. and A. 2007.S. 52 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . 2010. Optimal Placement of Phasor Measurement Units for State Estimation. Report (Washington. September 2011.” Public Utilities Fortnightly. DC. November 9–12.” adopted by the North American Electric Reliability Corporation Board of Trustees on February 10. and D. Maguire et al.” IEEE Transactions on HTS Power Cable to Be Installed in the Long Smart Grid 1 (2010): 11–19. Dynamic 37 of Long Island Transmission Level HTS Cable. Novosol. “Installation and Testing Results M.” Monitoring and Decision Systems Maximize IEEE Transactions on Applied Superconductivity 19 Energy Resources. forthoming. & Energy Magazine 8 (2010): 37–45. K. Charleston. NC. Superconducting Engineering Society General Meeting. 2009).” Journal of Regulatory 30 U. “Wide Area Measurements for Improved Protection Systems. 39 29 U. and R. Ott. Ilic..S. “ISO. Oren. April 16. 2724–2731. Novosel. (2009): 1692–1697. 2008. 34 Electric Power Research Institute. Phadke and D. O’Neill. “Standard PER-005-1: System Personnel Training. technical report 1015511 (Palo Alto.” Megawatt Daily. 33 M. NERC Pact Aims to Lower S. CA: 2009). technical report 1017793 Volume 3. and 27 D. March 15–18. Philbrick. “A Mixed 38 (2011): 961–966. AZ: Power Systems Engineering Research Center. Hrvatska. R. 2005). F. Transmission Switching. V. Generation Unit Commitment Problem. ed. O’Neill. Madani. WA. F. Smart Grid System Economics. 24 J.org/computing/software/ grid-visualization-efforts-helped-heal-after- hurricane/?utm_source=techalert&utm_ medium=email&utm_campaign=092911.” IEEE Transactions on Applied Superconductivity 21 D. “Analyzing Valid Inequalities of the Blackout Threat. 2009. http://spectrum. Revitalizing Operational Reliability. “Smart Transmission Grid Applications and Their 25 J. 2009. IEEE (New York: IEEE. Phasor Measurement Unit Implementation and Applications. Hedman.

electricity production uncertain because they are subject to only show a substantial increase from wind and solar limited control and the energy they produce is generation.1 failures of generation units. Variability and and renewable portfolio standards. moving generation scheduling decisions closer to real time. Section 3. We find that these operational changes will become increasingly important as VER penetrations grow.2 by a discussion of the impact that high penetrations of VERs could have on system operating reserve requirements and several ways to limit system operation cost increases.5 gives our conclusions and recommendations. For example. as well as context for later chapters. First. we recommend more widespread sharing of granular meteorological data measured at VER sites. we discuss the challenges of operating the grid with high penetrations of wind and solar generation. The chapter provides background on ongoing industry and government efforts to integrate and increase the penetration of VERs. such as demand response and energy storage.3 discusses the impact of high VER penetrations on the future well-adapted generation mix and the need to ensure adequate system flexibility. Section 3. Through decades of experience. system (EIA) expects renewable sources to constitute operators have developed approaches to cope 25% of the increase in total generating capacity with variability and uncertainty that stem.1 describes the basic technical and economic characteristics of VERs and introduces the challenges for the power system derived from the variability and uncertainty of these generation sources. we recommend that mechanisms that provide incentives for investment in flexible generation and for operating flexibly be devised and deployed in regions with growing VER penetrations. Most projections of U. generators are known to be variable and Chapter 3: Integration of Variable Energy Resources 53 .S. commonly referred to as “variable energy resources” (VERs). We also present a variety of changes to system operation and planning that can help in meeting these challenges. from changing demand levels and and 2030. We describe sources of system flexibility. Second. for both VERs and conventional generation technologies. including conventional generation technologies and potential new resources.4 discusses the critical role of interconnection standards in assuring that reliability is maintained as the penetration of VERs increases. Energy Information Administration systems. These standards. uncertainty are familiar concepts in power the U. Finally. It is particularly important that they be structured in response to anticipated rather than existing conditions. will need to adapt to the increasing role of VERs.Chapter 3: Integration of Variable Energy Resources In this chapter. This is followed in Section 3. wind and solar power of the power system.S. Section 3. and expanding cooperation among neighboring balancing areas. The topics described in this chapter are of primary interest to industry decision makers. But wind and solar generation as new sources of variability and Commonly referred to as “variable energy uncertainty present challenges to the operation resources” (VER). we recommend several changes to system operations that could facilitate the integration of VERs in many regions. which are receiving a strong push less predictable compared to energy from from state and federal policy through subsidies conventional technologies. for across the electric power sector between 2010 instance. Section 3. These include improving VER forecasts and situational awareness.

2007). Source: GE Energy Consulting. 2010 Cents Figure 3. Impact of Intermittent Generation on Operation of California Power Grid” (Sacramento. “Intermittency Analysis Project: Appendix B. The thicker lines indicate the monthly average profiles for load and wind.1 Daily Load and Wind Generation Profiles in the California System Compared to Average Profiles for the Month of January 2002 Average Load Average Wind Wind (MW) Load (MW) Hour Note: This figure compares daily wind generation and load profiles in California in January 2002. CA: California Energy Commission.2 Nevada Solar Photovoltaic (PV) Plant Output on a Sunny Day (Left) and a Partly Cloudy Day (Right) in 2008 PV Output – 5/1/2008 PV Output – 3/25/2008 50 50 50 50 MW MW Source: North American Electric Reliability Corporation. 54 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID Average Load . The colored lines in the upper portion of the plot illustrate each individual day’s load profile while the lines in the lower half of the plot illustrate each individual day’s wind generation profile. 2009). Figure 3. NJ. Accommodating High Levels of Variable Generation (Princeton.

Fossil fuel plants specific geographic area known as a “balancing must bear the expense of additional start-ups and region.2. Finally. However. and Figure 3. as well as operation at output levels diversity in VER resource availability and load distant from points of optimal fuel efficiency and over wide regions.1 CHARACTERISTICS OF VARIABLE The power system must have enough response capacity. an effort to balance generation with load over a incurring operational costs. large- seen here. and ramping can increase balancing areas can reap the full benefits of mechanical stress on generation plants. which drives an increase challenges associated with wind generation. In addition. espe- if the VERs ramp down or decrease production cially at the individual plant level. solar outputs are generally of time. cited as the main source of costs of integrating VER generators into the grid. poten- resource diversity by consolidating or cooper- tially resulting in higher maintenance costs ating with each other. illustrates some of the with low penetrations. and it does wide-area cloud cover. more frequent minute-by-minute variability. As illustrated in Figure 3.1. The power solar photovoltaic (PV) plants under cloud cover system must have enough response capacity. such as wind storms and and over a wider range than demand. and reduced life. and its relation to demand. shut-downs. the absence or very without storage can also produce outputs that are limited presence of the VERs during extended considerably more variable and less predictable periods of time (e. As in costs. create challenging not conform to daily cyclic patterns. Because ramp events are often difficult to forecast ahead Compared to wind. backup supply to maintain reliability standards under these worst-case conditions. and much as 90% over the course of seconds. individual also can result in operation challenges. smoothing out the effects of air pollution control. solar generation if the VERs ramp up. systems with control and electrical characteristics that are high penetrations of VERs typically must different from those of conventional commit more reserve generation than those Chapter 3: Integration of Variable Energy Resources 55 . An inverse operational situations known as “ramp events” correlation is sometimes observed in which wind that are characterized by a significant portion resources on average become most abundant of the VER fleet ramping up or down in unison during hours of limited demand at night. additional reserve generation (or considerably more cyclic and better correlated demand response) must be ready to increase with load. ENERGY RESOURCES from interconnections. dispatchable generators All the operational adaptations to accommodate or loads must make corresponding changes in VERs require more flexible power systems. typically achieving maximum output a production (or decrease load) to meet the load few hours before peak demand. demand response. storage.g.” Larger balancing regions have a greater shut-downs. demand response. wind output can change more rapidly scale weather events. the high uncertainty system stability requires careful planning of associated with VER outputs also can compli- investments.3 Since day-to-day Beyond the variability and uncertainty of wind speeds and cloud cover are significantly their outputs.2 over a period of minutes to hours. a plot of California’s wind generation backup supply to maintain reliability standards. several consecutive days) than wind. Adjacent smaller start-ups. storage.3. When wind variations or passing clouds result in changes in output. such as the timely addition of cate existing balancing processes and is often dispatchable plants with fast ramping capability. Exacerbating these concerns. wind and solar generation have less certain than the load forecast. from have been observed to vary their outputs by as interconnections.4 Providing these services at minimum cost while maintaining reliability and Beyond variability.

7 The consumers. Federal Trade Commission suggests that each the cost of reserves required to reliably operate VER plant purchase option contracts on flexible power systems has been allocated to all end resources that can provide reserves.9 Our focus is on wind farms and ting generation is typically expensive. and many have been maintain system stability and reliability under implemented in other countries and in parts the new anticipated conditions.5 This last scheme is in accordance with the As with any zero-sum game. to date VERs have been required penetration of VERs is also a subject of current to meet few performance standards in the U. debate.S. The allocation of the increase in system operat- some suggest possible implementations. In a November 2010 Notice of Bonneville Power Administration is exploring Proposed Rulemaking. Because retrofit. the Federal Energy the possibility of allowing wind generators to Regulatory Commission (FERC) suggested that pay the cost of their variability by self-supplying VERs be made partly or wholly responsible for the extra electricity to correct any imbalance.6 The (VER) is a subject of current debate. usually residential VERs historically have affected the U. BOX 3.8 the additional costs that they cause. including have been identified and studied in the wind VERs. it will become increasingly important that interconnection All the adaptations described in this chapter standards require all generators. synchronous generators. generators are lenges of system adaptation and the ways that initially dispatched according to unit commit- grid management and operation can help ment schedules made one day in advance. accounted for only a small fraction of energy supply. it will be difficult to design of balancing markets already in place in arrive at a mutually agreeable solution. Historically. This is exemplified by the responses to the notice. bulk and commercial rooftop PV installations. as described in Box 3. anticipated rather than existing conditions.1 ALLOCATION OF SYSTEM The respondents to the notice agree that such OPERATING COSTS a cost-allocation scheme for operating reserves will be hard to realize in practice.S.1. However. operated. implications of large-scale VERs for planning both VER and conventional. This chapter explores these chal. As explained in Chapter 2. As VER penetrations grow.2 VARIABLE ENERGY RESOURCES policy-backed technologies will require changes AND THE COST OF RESERVES to how systems are planned. in power system very little because they have Chapter 5. to play an active role in helping to integration literature. of the U. these 3. some European countries. The allocation of the increase determines which generators will come online 56 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . it is large-scale PV or concentrated solar thermal particularly important that interconnection installations without storage. Regardless. be designed for and building transmission systems in Chapter 4. with significant growth. larly now that the incurred costs are becoming significant. Due to their very low in system operating costs that results from high penetrations. and we discuss the distribution system challenges of distributed generation.S. FERC ing costs that results from the introduction of proposes adding a special ancillary services rate large penetrations of variable energy resources to the transmission access tariffs for VERs. We address the requirements for all generation technologies. particu. The minimize the potential grid-related costs of unit commitment or day-ahead market process VER expansion. and controlled.

system operators used during normal operation to maintain utilize generators’ bids instead of actual costs the balance between supply and demand due in determining how units are dispatched. and are rarely Commission (FERC) has cited this increase as explicitly dispatched by the operator. subject to transmission line capacity limits and associated with higher penetrations of wind will other security constraints.OAD FOLLOWING 2ESERVES MINUTES  4HESE ARE meet their dispatch schedule due to unforeseen dispatched during normal operations by the equipment failures or other contingencies. provisions must be made for the supply of operating reserves and balancing services. generators with lower marginal operating costs before those with higher costs.g. by dispatching increase operating reserve requirements. s 2AMPING 2ESERVE MINUTES TO HOURS  4HESE reserves respond to ramp events that occur Operators make these scheduling decisions over periods of minutes to hours. daily. a slower timescale than regulating reserves. ries of reserves are significant. s &REQUENCY 2ESPONSE 2ESERVES MILLISECONDS Wind integration studies and experience have to seconds): The fastest reserves are used to shown that the additional variability and respond to such contingencies as the loss of uncertainty associated with higher penetrations a generator or transmission line.” Either that are faster than the clearing periods of by central decision-making or through markets. operator to maintain the balance between necessitating the dispatch of reserves.10 The Federal Energy Regulatory tors and at control centers. or an uncommitted unit might be needed. such as wind forecast errors. a committed generator may not conditions.11 Chapter 3: Integration of Variable Energy Resources 57 . is known as They are dispatched by the operator in times “security-constrained economic dispatch. The committed generators may fail to s . timescale. They are of wind will increase operating reserve require- activated automatically on individual genera- ments. and resources are deployed during contingencies direction (up or down) of response expected. alongside faster reserves.or go off-line at various times during the Wind integration studies and experience have shown following day. They can under a level of uncertainty. They are designed In the context of VER integration. energy markets. be needed. discussed in Appendix B. This to random changes in generation or load. process. tors and demand response can provide these weekly) changes in demand or generation. Genera- supply and demand due to cyclical (e. five catego- to slowly ramp up and replace faster reserves. Vertically integrated utilities that the additional variability and uncertainty attempt to minimize the cost of meeting load. Where there are s 2EGULATING 2ESERVES SECONDS  4HESE ARE organized wholesale markets. one of the most important sources of cost i A variety of terms are commonly used to describe each type of reserve in different systems and different countries. on operating reserves. on the day are designed to cover challenging operational of dispatch. listed in order which are then available to address future of response time: i contingency events. Operating reserves are categorized according to s 3UPPLEMENTAL 2ESERVES TENS OF MINUTES TO the types of events to which they are designed hours): The slowest form of reserves. Since forecasts are ramp their outputs either up or down and never perfect one day in advance. these to respond and to the speed.

3 The Evolution of Forecasting Errors versus Lead Time. iii An exception is the occurrence of wind exceeding maximum allowable speed and causing multiple turbines to shut down simultaneously. Ramping and load-following reserves are Accommodating these fluctuations may require primarily used to counteract VER forecast a modest increase in fast-responding regulating errors in the day-ahead scheduling of plants. 2005–2008 Mean absolute error / mean production (%) Hours Note: The red bars mark the typical time when generation schedules become final and binding for different markets. concluded that large VER penetrations do not as illustrated in Figure 3. At this timescale.15 to be comprised of many small plants on the As a result.S. and it is highly unlikely that VER production could be at least 20% lower ii These studies assume that requirements for fault tolerance are present in grid interconnection standards.4. Integrating Intermittent Renewables Sources into the EU Electricity System by 2020: Challenges and Solutions (Brussels. operators must conservatively order of 1–5 megawatts (MW) over wide operate the system assuming that the actual geographic areas. they typically take minutes require a closer examination of how VERs to hours. significantly increase the risks for traditional day-ahead load forecast errors are typically contingencies.3.14 For comparison. 58 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .12 Understanding the full impact of this stop producing. The uncertainty and the variability of VERs can create fluctuations Ramping and Load-following Reserves. This issue is described further in Section 3.13 Even when VERs do simulta- increase in operating reserves on costs will neously ramp down.ii Large penetrations are expected below 1% mean average error of production. reserves required for normal operations. wind forecasts have an average error of 15%–30% mean absolute error Frequency Response and Supplemental of production. Figure 3. in production on the order of minutes. Belgium. which is much slower than traditional affect each category of reserves. contingencies. increases from the integration of VERs in the many of these plants simultaneously would U. Source: Eurelectric. 2010). All major international studies have ments in wind forecasting over the past decade. that occur in seconds. despite the significant improve- Reserves. such as a large generation plant tripping off-line.iii Regulating Reserves.

2011). Wind reserves associated with VERs. The blue line indicates actual wind generation.4 Ramp-up and Ramp-down Events in the Bonneville Power Administration Region Note: The red line indicates the final scheduled wind generation (economic dispatch base point). which are then dispatched throughout portion of the agency’s total wind generation the day if forecast errors become apparent.000 MW—a substantial reserves. NJ.16 operator responses. requiring large quantities of December 7 to nearly 2.4 Report: Operation Practices. The risk of ramp events in most cases then ramped down to zero through the night comes from the uncertainty of when they will hours of December 7 and the morning hours occur and how long they will last. than forecasted.000 MW at the time). we discuss a reserves procured by the Bonneville Power set of tools that can serve to reduce the cost of Administration in December 2009.Figure 3. Source: North American Energy Reliability Corporation. Many of these output ramped up sharply in the late hours tools will allow the electricity market to better of December 6 and the early morning of address the impending need for flexible Chapter 3: Integration of Variable Energy Resources 59 . NERC IVGTF Task 2. Wind output concern. some wind generation by issuing generation Though not a traditional contingency. BPA deployed more than 90 percent of its reserves Load-following and ramping reserves also are as the wind ramped up and BPA had to curtail used to hedge the risk of VER ramp events. Procedures. in order to avoid events can present a significant operating violating its reserve requirements.4 shows two ramp events that depleted In the remainder of this section. of December 8. and Tools (Princeton. ramp limits to wind generators. This down ramp resulted in predictions of ramp event occurrence and BPA depleting 100 percent of its reserves and timing are often difficult and can result in large prompted BPA to curtail transmission schedules and sudden forecast errors that deplete reserves for wind generators. capacity (around 3. Figure 3. Indeed. Uncertainty in both the that are on-line in a time too short to activate size and duration of ramp events complicates reserves that are fully off-line.

. the Local meteorological and power production fastest flexible resources will be able to offer measurements at individual VER installations their services closer to the time when they are are also an important input to VER forecasting actually needed.21 Increased data sharing would improve power Wind power generation forecasts used in power production forecasting. and they are rarely refreshed slowly and have relatively poor spatial shared beyond forecast vendors and system resolution. with significant penetrations of VERs. 60 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . and the agency is Of course. resources: if the electricity commodity market is liquid and a strong price signal exists. this is not universally the case. the atmosphere at a given time. Improvements in the affordability operators. the National Oceanic of current NWP models and has urged FERC and Atmospheric Administration (NOAA) to consider also mandating VERs to share local recently began updating NWP data every hour meteorological measurements with NOAA. This allows the system to avoid models.20 In a November 2010 Notice hedge this risk. (ISOs) and regional transmission organizations tion forecasts is one of the most straightforward (RTOs) exist are required to provide data to ways to reduce the impact of VERs on reserve system operators for the purpose of forecasting. they are typically etary and confidential. In the U. requirements and thus system operating costs. rological and power production data measured and these increases may exceed the benefits in at individual VER installations. protections would need to be in developing further improvements. Reducing real-time forecast of Proposed Rulemaking. These datasets are typically refreshed locking in resources that it will not eventually more often than the NWP data that NOAA need. a final describe many aspects related to the state of rule had not been issued. weather prediction (NWP) models maintained increased data reporting and compliance by public meteorological agencies and meteo. However. As of November 2011.S. and these can be freed for other uses. monitoring will also result in increased costs. diminishing the imbalances of centralized forecasting in real time or that must be corrected by regulating and load. FERC proposed to errors can increase the accuracy of the real-time mandate reporting of such data for the purpose dispatch schedule.18 place for commercially sensitive information.19 Reducing day-ahead forecast errors lessens the System operators have indicated that accurate risk of under committing generation. due to the computational complexity of the models Local data from VERs are considered propri- used to generate these datasets.17 provides and have higher spatial resolution. thus also data from VERs are critical to generating diminishing the ramping reserves needed to accurate forecasts. However. At the same time. VER generators in some regions where independent system operators Improving the accuracy of VER power produc. and is widely under- system operations typically are based on a stood to be an important step in reducing the combination of data from large-scale numerical cost of VER integration. near-real time by system operators in regions following reserves. NWP datasets regions that do not expect substantial penetra- are wide in regional scope and comprehensively tions of VERs.23 instead of every six hours.22 NOAA has cited limited access to and power of computational technologies have wind data at wind turbine–hub height as a yielded significant advances in NWP models in significant limitation to the prediction accuracy recent years. Individual VER owners and forecast vendors Improve Variable Energy Resource Forecasts collect and use the data to produce power and Situational Awareness production forecasts.

to take advantage of this phenomenon: shorter developing and improving tools for ramp-event gate-closure periods (the time before the prediction also can reduce the impact of VERs operating period when generation schedules on system operating costs. as shown in Figure 3.3. a forecasting ment periods (the length of time the generator system for ramp events has the goal of identi- is committed to produce). While these markets phasor measurement units.28 increased data access may yield.27 still buy or sell in shorter-term markets that are closer to real time. would improve wind forecast accuracy in the accuracy of wind forecasts increases appre- regions with high penetrations of VERs. are described in differ by name and implementation. to counter the variability and uncertainty of System operators can take specific preventative loads. it has been operate real-time markets that layer on top suggested that providing forecasters with of day-ahead markets. Intraday markets have been Reducing the potential impacts of VER ramp developed on top of power exchange markets. Finally. in practice Chapter 2.24 Make More Frequent Decisions Closer FINDING to Real Time Greater sharing of meteorological data As discussed earlier and illustrated in Figure 3. Like the case for Notably in the U.29 observe these conditions. A generator with an greater access to meteorological data would obligation from the day-ahead market may improve forecasts. to increase operators’ situational awareness of and their use has been mandated recently system conditions.. events also will require deploying technology initially in Spain and then in other countries. Chapter 3: Integration of Variable Energy Resources 61 . Historically. most ISO/RTO regions improving regular VER forecasts. Unlike a conven- become final and binding) and shorter commit- tional VER forecasting system. data analysis and they all shorten the gate-closure period. There are two primary mechanisms for improving scheduling practices Beyond improving power production forecasts.Such confidentiality arrangements already exist visualization tools that effectively aid operators’ between NOAA and many airline companies.S. wind-forecasting companies. both fying specific meteorological conditions that measures have been deployed in various regions could translate to high risks for ramp events. Some of these tools. Recent research and operating experi- actions to protect system reliability and time- ences from around the world have shown that lier. such as within the European Union.S.25 A recent review of ramp event-forecasting methodologies by Moving gate closure nearer to the corre- Argonne National Laboratory indicates that sponding operating period results in more ramp-event prediction remains a relatively accurate forecasts. Many NOAA also has been working to form volun. and novel problem for meteorologists and reports better-performing schedules in systems with that existing forecasts tend to be “unreliable a significant penetration of VERs. decision-making will be important. less resource-intensive decisions to address they also can yield potentially large benefits in potential ramp events when forecasting tools accommodating large penetrations of VERs. driven both gate-closure periods due to the large benefits by public research institutes as well as private that are available even in systems without VERs.” 26 Fortunately. ramp- mechanisms already exist to achieve shorter forecast skill is rapidly increasing. with growing tary data-sharing partnerships with industry penetrations of VERs are in the process of stakeholders and has embarked on efforts to developing programs to aid in identifying and better quantify the potential benefits that responding to the risks of VER ramp events. ciably closer to real time. Various and of low accuracy.3. system operators in the U.

the electric grid is managed by transmission owners that Cooperate among Balancing Areas rely mostly. Germany.3 – 15. most RTO/ISO regions already points in mind.31 This arrangement can result in roll over more smoothly and forecasters to discrimination between vertically integrated refresh their forecasts more often. With increasing penetrations Implementing shorter gate-closure periods 50 50 50 50 of VERs.5 Individual and Aggregated Power Outputs of Wind Turbines Single Turbine Oevenum/Föhr) 225 kW Normalized Power Group of Wind Farms (UW Krempel) 72. research note 2493 (Espoo. With these In the U. the overall cost of maintaining real-time which most often vary over tens of minutes. on bilateral transac- tions between market participants. FERC proposed to gate-closure periods and five. the portions of the Western Interconnection a final rule has not been issued.. which reduces utilities who own the transmission network the magnitude of forecast errors and associated and independent generators who own VERs. if not entirely. Finland: VTT. Phase One 2006–2008.9 GW Source: Copyright © Fraunhofer IWES.30 balance between generation and load in systems with high penetrations of VERs.S.to fifteen-minute reduce scheduling intervals nationally from scheduling and dispatch. shorter operating periods also could or shorter scheduling periods in these regions align market operations more closely with the would help mitigate such a concern and reduce predominant variations associated with wind. 2010 Cents Average Load Average Wind Shorter operating periods allow schedules to in advance. Holtinnen et al. however. H. For the Southeast and 1 hour to 15 minutes. Design and Operation of Power Systems with Large Amounts of Wind Power: Final Report. Generators Enlarging the regional scope of power system selling into these systems usually are required operations through cooperation or integration to submit fixed hourly schedules that are set among balancing areas can offer reliability 62 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . 2009). Figure 3.33 As of October 2011.. Institute for Solar Energy Technology. Wind Energy Report Germany 2005 (Kassel. 2005).7 MW All Wind Turbines in Germany 14.32 scheduling errors. outside the California ISO. Germany. in a November 2010 Notice administer wholesale energy markets with short of Proposed Rulemaking.

New York. the creation savings across the WestConnect region of of wide-area balancing markets. and the Bonneville wide-area aggregation requires that sufficient Power Administration has several pilot transmission be available. jurisdiction to another. Mississippi. Nebraska. it are discussed in Chapter 4. amounts of VERs. a trend However. rapid changes Kansas. and associated increase in variations of the aggregated market liquidities made cooperation among resource.36 for the explicit purpose of VER integration remains a relatively new activity. decades before VERs and economic benefits when integrating large became widespread. The changes in the outputs of individual resulting decrease in variability. One study found the day-ahead WestConnect created the Virtual Control Area forecast error could be reduced by as much as Work Group to investigate methods and a factor of two when the geographic region technology available for participating balancing diameter was increased from 140 kilometers areas to function as a single “virtual” control to 730 kilometers. Texas.40 Ultimately.35 The cause is the geographical and the PJM interconnection all grew out of the diversity in the weather that is seen by consolidation of independent balancing areas geographically dispersed turbines and plants. cooperation among balancing areas that is similar for solar generation. Figure 3. within their respective regions. Utilities in Arkansas.39 For example. including the interregional commu- and Solar Integration Study identified cost nication of regional imbalances. Regions like the by the slower output variations of the aggre. and New England ISOs gated resource. balancing areas economically worthwhile in many parts of the U.34 By aggregating a geographi.37 dynamic scheduling between balancing areas. However.5 illustrates the impact on variability of spatial dispersion of wind resources. and interhourly or assuming that adequate transmission exists. and New Mexico cooperate via in the outputs of individual VERs are replaced the Southwest Power Pool. Louisiana. Chapter 3: Integration of Variable Energy Resources 63 . A variety of The economic benefits of cooperation between possible cooperation schemes have been regions could be significant: the Western Wind proposed.S. the forecasted aggregate Corporation (NERC) has recognized the need output of turbines over a wide geographic area to review and study interregional cooperation. rapid historically have been observed with the aggregation of loads over wide regions. this issue has not been studied in depth in the U. of $43 billion if the area operated as five large the transfer of load responsibilities from one regions rather than many small zones.S.By aggregating a geographically The benefits of geographical smoothing diverse collection of VERs. given the programs in place. is unlikely without new construction. Oklahoma. Midwest. the provision $2 billion from a total annual operating cost of operating reserves from other jurisdictions. cally diverse collection of VERs. which.. balancing areas will determine the degree to This issue and recommendations for addressing which cooperation or consolidation is possible. turbines. balancing VERs are replaced by the slower output responsibilities.38 Capturing the benefit of area for specific operations. is more accurate than the forecasts for single and industry-led efforts exist. Limited transmission between balancing areas may slow their aggregation. The North American Electric Reliability By the same principle.and solar-rich sion capacity that exists between neighboring locations. the transmis- remoteness of many wind.

however. substantial changes in output can take hours to tens of hours. and large coal-fired generators. while hydro units wholesale markets may have insufficient invest- can ramp in fractions of a minute. These plants.3 ENSURING ADEQUATE SYSTEM FLEXIBILITY The increased demand for flexibility from VERs will not necessarily translate into a need for The operation of a system with a substantial new capital investments.42 In preliminary studies of VER the level of VER penetration. FINDING negative economic consequences.43 Flexibility is not a new concept in power systems. the rate at resource penetration anticipated during which a generator can change its output. several presence of VERs will be different from today’s regions already experiencing high penetrations operation. reduced life. sufficient flexible capacity to accommodate modate the variation in the output of the VERs. These flexible integration in North America. offer reliability and economic benefits and the financial compensation paid to when integrating growing variable energy generators for their start-up costs may become resource penetrations. can effect changes in their output increase. or “cycling. in order to remain economically viable. or both. the anticipated growth in VERs over the next decade or two. already have modifying their output. although this is not known as its ramp rate. is limited by its design and technology.S. Other units. such as gas As expected VER penetrations continue to turbines. the next decade. its importance has been FINDING greatly amplified in recent years by the prospect The existing generation fleets in some of large penetrations of wind generation. further increasing the financial incentives they would need to cycle regularly. Some units. too. The future well-adapted mix of of wind power have had enough flexibility in generation technologies also will change. and expanding designed for frequent cycling will incur cooperation among neighboring balancing increased maintenance costs. regions with organized within fractions of an hour. But ment in flexible plants due to the uncertainty requiring resources to cycle frequently has regarding the most appropriate technology mix. In Europe. need to operate with high capacity factors these changes. 3. Large base- Improving forecasts. their varying output moves generators areas are operational changes that can away from their point of maximum efficiency. Technically. some regions resources must be capable of continuously have concluded that they.41 regions appear to have sufficient flexibility Power system flexibility has both technical and to accommodate levels of variable energy economic components. characterized by and utilities in many regions in the U. System operators significant. however. have slow ramp rates.” to accom. are relatively high capital costs but low variable making progress toward implementing costs. reducing gate-closure load units and midrange plants that were not and operating periods. 64 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . always depending on investment. their existing generation fleets to successfully probably reducing the proportion of less flexible integrate significant levels of wind without baseload units and increasing the percentage of substantial new dispatchable generation more flexible resources. such as nuclear universally the case.

If necessary. fuel costs and lower efficiencies. and some particularly of slower. including combined-cycle gas turbines.47 Chapter 3: Integration of Variable Energy Resources 65 . intermediate. or baseload combined-cycle gas generation. including diesel and gas combustion turbines.44 There is no consensus yet on the most appropriate solu.45 ramp rates. including nuclear. gas. experience has shown that the factors constraining their flexibility are mostly economic. nuclear—account for around 90% of the ments. Thermal Generation tions. ensure system reliability and efficiency. incentives to ensure that flexible resources they can be costly to operate because of higher will offer their flexibility to the power system. in flexibility and operate flexibly within the system.the rate of renewables growth. With key transmis- future prices and operating conditions. their usefulness may be limited if economics of such a mix under anticipated transmission constraints exist. While they have the lowest to the markets. In thermal plant and operating it flexibly require iv It is technically possible to operate nuclear generation facilities flexibly. generation fleet. they generally have topic in the next stage of its Grid Integration of relatively narrow power output ranges and slow Variable Renewables project. Agency found that more than the incentive improve on the efficiency of peaking plants. or other regulatory instru. requiring the pated penetrations of wind generation are dispatch of more expensive resources. Another important issue is offering sufficient can be quickly started and ramped. the most economic flexible European countries with significant or antici.iv The optimal mix of the three categories Generators require strong and clear of generation should consist of a portfolio with an adequate balance of flexibility and economic incentives to make investments production costs. Achieving flexible operations requires relatively minor technical modifications to plant designs and fuel content. to offer the full extent of their flexibility operated flexibly. are not usually plants. Baseload needed to prompt owners of flexible resources. such as operating reserves. we describe several existing and thermal generator flexible and accessing this upcoming resources capable of offering flexibility flexibility incurs costs. plants. Électricité de France regularly cycles their reactors to provide a range of grid services. Some of their units experience up to 250 power variations between 10% and 80% of their nominal rated capacity per year. Designing a flexible to systems with high VER penetrations. appropriate regulatory capacity of the U. currently addressing this issue. which could include enhanced capacity mechanisms.S. including primary frequency regulation and daily load following. and ancillary services. for this purpose. resources can become inaccessible. Research is needed to design While some technological limits exist in the market rules and incentive mechanisms flexibility available from thermal generation. Peaking plants. However. provided by fluctuating electricity prices will be losing some flexibility in the process. The agency plans to address this per-kWh operating cost. Making a In this section. coal. and the practice. Several sion lines congested. Intermediate A recent study by the International Energy plants. new categories of remunerated Thermal generators—mainly coal.46 They measures or changes to market design should supply the bulk of the flexibility within the be developed to facilitate adequate levels of power system today by participating in energy investment in flexible generation plants to markets and offering dispatchable capacities. and they incur significant real and opportunity costs when made to ramp and FINDING cycle.

51 As a tion in this region. the operation of hydro generation that all decrease the efficiency of the system. when VER production levels are low and marginal production costs are high enough. they can store potential has been identified from small hydro and the energy for later generation by varying the conversion of non-powered dams to powered water levels of their reservoirs. a number of redesigns and operating practices In practice. mally high or low water.Constrained by the need to comply with emissions Hydro is also an economical form of generation regulations. even though it is facilities are able to rapidly ramp their output highly viable as a generation technology. “hydro generation” refers to impoundment. some thermal generators may require because it can operate with essentially zero fuel cost. This loss of flexibility coincided with a period Hydro Generation of high wind. especially in the case of coal along the Columbia River during the summer generation.S. ones. forcing wind generation to be involuntarily curtailed. as many VERs can offset hydro production when their of them may end up being built with limited production levels are high and allow water to water storage capacity. particularly during times of abnor- carbon dioxide. although water used for generation today additional emissions-reduction investments before will not be available if needed in the future. pure run-of-river installations are typically smaller and have less impact on the environment. The water generation for the purpose of providing system collected then can be used to generate power flexibility is likely to be valuable but limited. makes up Hydro generation has two characteristics that only about 8% of total capacity and supplies make it a highly flexible resource.52 over a wide range while maintaining high However. these facilities produce variable outputs and are often categorized as a type of VER. cycling fossil fuel–based and businesses. for generators will increase maintenance costs and example on fish migrations. some potential for capacity expansion efficiency. The high water flows or particulates. incurring an opportunity cost. decrease equipment reliability and life. Second. tourism.49 result. v In this report. water quality. However.54 66 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . sulfur oxides.53 It is unclear how much grid flexibility would be obtainable from these plants.48 Constrained by the need to of 2010 limited the flexibility of the Bonneville comply with emissions regulations. or dam-based. hydro generation in the U. continuously cycling thermal flooding and other adverse impacts. nitrogen oxides. as thermal generators may require additional the plants were forced to operate in order to emissions-reduction investments before their minimize water spilling over the dams and keep flexibility can be accessed. total nitrogen saturation at levels safe for fish.50 The success of wind in Denmark and its integration in the Nordic countries is generally Concerns over flooding and environmental attributed to the availability of good intercon- impacts limit not just operation but the nections and large quantities of hydro genera- construction of hydropower facilities. These limits can restrict their generation also may increase emissions of flexibility. This is in contrast to run-of-river hydropower installations with little or no water storage. By avoiding the high dams and extensive land inundation that characterize impoundment hydropower. hydropower unless otherwise noted. The future role of hydro accumulate in hydro reservoirs.v First. their flexibility can be accessed. particularly in Norway. and nearby residents reducing efficiency. is usually subject to limits in place to minimize Furthermore. some Power Administration’s hydro generation. By aquatic ecology. these around 7% of total energy.

electricity use in response to signals from the system operator or changes in the price of A challenge of expanding traditional load- electricity. and merits further study. pricing programs that provide customers with Emergency and interruptible load programs greater transparency and control may offer provide customers with incentive payments or attractive alternatives to load control programs. discussed in detail in potential to offset operation and capital investments Chapter 7. This technique already has been Energy storage using pumped water. rate discounts in exchange for load reductions particularly if combined with automated during declared system emergencies. time-varying prices may response for higher-frequency adjustments also induce load shifting that makes better use of requires a better understanding of the speed off-peak VER generation. This flexible generation. providing real-time flexibility on a the future. this and pump. and predictability of the response of load to control signals and the behavior of end users in The earliest and most widely adopted applica. event in the Electric Reliability Council hydro energy storage (PHES). batteries. During periods of excess energy from by avoiding the need to hold additional reserve VERs.760 hours of the year to deal with relatively rare ramp Faced with the economic trade-offs events. Dynamic flexibility is in response to reliability threats.55 As discussed in the ator and water turbine can operate as a motor Western Wind and Solar Integration Study. response represents a small but growing share demand response and energy storage could play of operating reserve and regulation service important roles as new sources of flexibility in markets. control demand response programs to accom- demand response that is sufficiently nimble and modate VERs is the concern that more frequent reliable has the potential to offset operation activation may lead to fatigue and an eventual and capital investments of more expensive lack of willingness of load to participate. including a February 26. and other storage tech- VER ramp events. unusually high demand or loss of a major dynamic pricing is in its infancy in U. elec- generator or transmission link threatens a tricity markets.Additional Sources of Flexibility generation online for all 8. flywheels. such as Georgia Power’s real-time demand response may be used to maintain pricing tariff for large industrial and commer- stability by calling on customers to shed load cial customers. where the gener- of Texas (ERCOT). that is sufficiently nimble and reliable has the Demand response. occurs when customers modify their of more expensive flexible generation. an environment of rapid and perhaps frequent tion of demand response in the context of curtailment or load adjustment. As discussed in Chapter 7. is of particular value as a flexible application of demand response reduces costs resource. suggest its efficacy. compressed extended to protect system reliability during air. as often is pilots. or manually disconnecting specific customers from the grid. Because demand response works more routine basis. In the context of VER integration. it has the potential to offer flexibility at lower marginal In the context of VER integration. When control technology.57 Pumped 2008.56 In some regions with wholesale surrounding flexible thermal generation and markets—most notably ERCOT—demand the scarcity of resources for hydro generation. Use of demand true of on-shore wind. nologies could also supply flexibility. programs. Where VER output is not has not been targeted in demand response well correlated with load patterns. by offsetting physical generation. although successful long-term power system’s operating reserve margin. water can be pumped into the elevated Chapter 3: Integration of Variable Energy Resources 67 . demand response costs and emissions than thermal generation.S.

reservoir and used to generate at a time when To date, utility-scale battery and flywheel
rapid up-ramping is required, providing time for technologies have achieved limited energy
slower units to respond. The U.S. currently has capacity and have only been deployed in a small
approximately 22,000 MW of PHES capacity.58 number of pilot projects.62 These technologies
remain too costly for most applications today,
PHES facilities do face important limitations. at a price about two to five times higher than
They are only viable in locations that have competing sources of flexibility.63 Storage
sufficient water availability and are capable of technologies are currently the focus of research,
siting large reservoirs at different heights. Many and costs could fall in the coming decades. If
of the best locations for PHES facilities in the they do, the use of bulk energy storage in the
U.S. have already been developed. Environ- power system could expand dramatically.
mental concerns regarding the construction
of large dams and reservoirs further restrict Obviously, curtailment of VER generation is
the viability of additional PHES installations. another source of flexibility. VER output can
Only one PHES facility with capacity of more be reduced by temporarily disconnecting
than 100 MW has been constructed in the past individual generators. Some advanced wind
15 years.59 FERC has recently issued preliminary turbine designs also allow controlled feathering
permits to more than 40 projects that total over of their blades to partially reduce their output.
32,000 MW of additional PHES capacity.60 While VERs do not appear to provide operating
However, a preliminary permit does not autho- reserves anywhere in the U.S. today, studies have
rize construction, and it is unclear how many of shown they are technically capable of doing so.64
these projects will eventually be constructed.
In some cases, curtailing VER outputs can
Compressed air energy storage (CAES) is the be the economically optimal decision—for
only other storage technology that has achieved example, during “minimum generation events”
long-term utility-scale operation. CAES facili- when the combined generation of wind and
ties use electricity to compress air and inject it baseload facilities exceeds the load. If the
into underground caverns for storage. When the baseload plants are already operating at their
energy is needed, the compressed air is heated technical minimums, either they must be shut
and run through a turbine to generate elec- down or the wind plants must be curtailed.
tricity. Only two utility-scale CAES facilities have This latter option is often more economically
been constructed worldwide: a 290 MW facility efficient in the short run because it avoids the
with two hours of storage in Huntorf, Germany, large costs associated with shutting down the
that entered commercial operations in 1978 and baseload plants and then starting them up a
a 110 MW facility with 26 hours of storage in short time later. In the long run, the prospect
McIntosh, Alabama, that entered commercial of baseload units being shut down during
operations in 1991. Supported in part by funds minimum generation events would provide
from the American Recovery and Reinvestment added incentive for investments in plants that
Act of 2009, New York State Electric & Gas and can operate more flexibly at lower cost.
Pacific Gas & Electric have recently announced Significant penetration of VERs in general will
plans to construct new CAES facilities with cause baseload generation facilities to cycle
respective storage capacities of 150 MW/16 more frequently, potentially leading to loss of
hours and 300 MW/10 hours.61 These efforts efficiency of these plants and potential emis-
promise to advance the maturity of CAES sions increases, though this issue is not yet well
technology and reduce the uncertainty understood.65 VER “must-take” operating
regarding its cost at commercial scale. practices, as they are currently established in
the EU, will amplify this effect.

68 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

FINDING Most wind turbines and all solar PV plants
Care must be taken in the design of lack significant inertial response. While wind
operating procedures to ensure that the turbines do store mechanical energy as rotating
option to curtail the output of VERs is always inertia, there is a significant distinction between
them and conventional generators. While the
available to operators and that it is selected
mechanical input power to conventional
when it is economically efficient to do so. generators can be increased to return the
generator to its correct speed, this cannot be
done for wind turbines as the mechanical input
3.4 INTERCONNECTING VARIABLE power is the wind, and this cannot be increased
ENERGY RESOURCES except in the rare case where the wind turbine
output has been curtailed.vi When wind and PV
Beyond the variability and uncertainty of their account for larger portions of the generation
outputs, VERs exhibit other characteristics that fleet, the inertial response of the overall system
are very different from conventional generators. will decrease, potentially increasing the risk of
Wind turbine generators do not operate at power system stability problems. Reduced
constant speed, which in most cases results in system inertia due to the displacement of
an ac output with variable frequency and
voltage that is not directly compatible with the When wind and PV account for larger portions of
grid. Solar PV systems produce a dc voltage, the generation fleet, the inertial response of the overall
which also cannot be directly connected to the
system will decrease, potentially increasing the risk
grid. Both require an interface system based on
power electronics that converts their outputs of power system stability problems.
to grid compatible form. conventional generation with wind already has
been observed in the Irish and the Hawaiian
The physical characteristics of VER generation systems, and researchers have raised a similar
and its specific connection to the grid result in concern for the Western Interconnection.66
a different contribution to the inertia of the It has been shown that wind generators can
power system than that of conventional plants. provide some degree of inertial response through
All conventional forms of generation produce appropriate design and control of their power
electricity by rotating a large metal mass within electronic interface, and solar PV generators
a magnetic field. Conventional generators also can emulate an inertial response if provided
include a rotating turbine with large mass. with energy storage.67 However, as shown by
The considerable stored kinetic energy of this experience, current wind turbines have an
aggregate mass gives these machines the inertial response performance below that of
transient ability briefly to provide electrical conventional generators.68
output power in excess of or below their
mechanical input power to accommodate Other characteristics of VERs also need consid-
a sudden change in load—in effect, to act like eration. Because of their power electronics–
a shock absorber for the system. Power systems based interface to the grid, VERs can inject
rely on this inertial response for maintaining harmonic voltages or currents into the grid.
system stability. During such a transient, the This may result in a distortion of the sinusoidal
mechanical input to the generator is changed to voltage waveform experienced by other pieces
return the generator speed to its proper value. of equipment connected to the system, poten-
tially disrupting their operation. The varying

vi
Some modern wind turbines can employ blade-pitch control to affect some control of input.

Chapter 3: Integration of Variable Energy Resources 69

power output and complex control systems of system voltage reduction, a capability known as
VERs could cause voltage to oscillate rapidly, or “low-voltage ride-through.” Power systems
flicker. Finally, because their response to faults employ sophisticated protective relaying and
is different from that of conventional genera- control schemes throughout the transmission
tors, VERs often require customized protection system to clear faults quickly when they occur.
equipment.69 When faults do occur, the system voltage often
drops until the fault is cleared. It is important
Most of these challenges can be readily for generators to remain connected during this
addressed by enforcing a set of interconnection time as a substantial loss of generation could
standards. The interconnection standard further reduce voltage and threaten the stability
provisions for conventional generators are of the system. The order also requires VERs to
designed to ensure that generators do not harm provide reactive power support, which allows
the grid and that they will contribute to the them to contribute to voltage stability, and it
stability and reliability of the grid when required. requires wind generators to be compliant with
Due to their very low penetrations, so far VERs the existing supervisory control and data
in the U.S. have been required to meet few acquisition systems that utilities and system
standards. As a result, their expected impact on operators use to remotely control and monitor
the system has not yet been properly formalized, the power system.
and they generally have not played an active role
in maintaining system stability and reliability. FERC has yet to issue a similar order for
solar-based generation technologies. If the
FINDING penetration of solar generators increases to a
Before the penetrations of VERs in the significant level, performance standards will
become important. Solar PV generators in
U.S. increase to those levels foreseeable
particular share many technical characteristics
with renewable portfolio standards,
with wind generators, so it may be appropriate
interconnection standards must be revised to extend the provisions of Order No. 661-A to
to include VERs within the power system them, as the Interconnection Standards Review
and allow them to perform functions that Initiative of the California ISO recently
can enhance grid behavior. These functions proposed.71 Doing so will require the resolution
must be compatible with their unique of inconsistencies between Order No. 661-A
and existing standards for solar generation
physical and electrical characteristics.
interconnection. For example, Order No. 661-A
requires wind generation to have low-voltage
The nation has made some progress over the
ride-through capability, while the primary
past decade on the development of intercon-
technical standard governing the interconnec-
nection standards. FERC issued Order No.
tion of distributed solar generation requires
661-A in December 2005, specifying that wind
distributed generation to disconnect immedi-
generators meet three performance requirements
ately upon sensing low-voltage conditions.vii
before interconnection rights are granted.70 The
NERC currently has two task forces designed to
first requires that wind generators not discon-
reconcile these standards and develop others
nect from the system in response to a transient
for distributed solar generation.72

vii
This provision is in the Institute of Electrical and Electronics Engineers Standard 1547.

70 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

Recent technical developments have enabled 3.5 CONCLUSIONS AND
engineers to construct wind generators that RECOMMENDATIONS
are compatible with most of the stability and
reliability performance requirements typically Motivated by state and federal policies, the
imposed upon conventional generation.73 proportion of generation from VERs in the U.S.
Manufacturers typically construct modern VER could grow substantially over the next several
plants with advanced power electronics and decades. While variability and uncertainty are
controls. Among other capabilities, these power familiar concepts in power systems, wind and
electronics give plants reactive power control. solar generation, as new sources of both, could
Advances in mechanical engineering also have complicate power system operations and
allowed modern wind turbines to curtail their planning.
power outputs with precision and in real time
through rotor pitch control. This capability, High penetrations of wind and solar generation
known as active power control, allows the will require operators to procure additional
turbines to maintain a fixed power output operating reserves. Increased operating reserve
below maximum or a maximum up-ramp rate. requirements are viewed as one of the primary
sources of potential system operations cost
NERC, ISOs, RTOs, and utilities are closely increases associated with the growth of VERs.
monitoring the impact of VERs on grid reli-
ability to determine whether existing require- Improving the accuracy of VER power produc-
ments should be modified or additional tion forecasts and VER ramp event forecasts
requirements are needed. Following recom- could reduce the magnitude of these cost
mendations from its 2009 report on integrating increases. Improving both types of forecasts
VERs, NERC has created a subgroup within its relies critically on forecasting methodological
Integration of Variable Generation Task Force improvements and the increased availability of
to review the adequacy and consistency of U.S. granular meteorological data, especially data
interconnection standards for all generation, measured at individual VERs.
both VERs and conventional.74 The task force
is expected to release a report by the end of R E CO M M E N D AT I O N
2011 making recommendations on intercon- Industry and government should work
nection issues.
to expand the sharing of granular
meteorological data measured at VER sites
FINDING
for the purpose of improving wind power
Interconnection requirements for all
production forecasts.
generation technologies, both variable
energy resource and conventional, must
be designed for anticipated rather than
existing conditions. The technology
necessary to comply with anticipated
standards is available.

Chapter 3: Integration of Variable Energy Resources 71

Changes to system operating practices could needed in other regions and further into the
also facilitate the cost-effective integration of future. This flexibility can be provided by
high penetrations of VERs. Integrating forecasts thermal or hydro generation, demand response,
fully into power system operations could yield and energy storage. But it is likely to prove
significant benefits and remains an important difficult to expand hydro capacity. There are
challenge, but one industry appears to be tackling concerns regarding the ability of present
well. Moving the deadline when generation wholesale market designs to attract the
schedules become final and binding closer to necessary volume of flexible resources.
real time and reducing the duration of genera-
tion commitment periods would allow system R E CO M M E N D AT I O N
operators to take advantage of the fact that As VER penetrations increase, mechanisms
wind forecast accuracy improves appreciably
that provide incentives for investment
closer to real time. These changes promise to
in flexible generation and for operating
reduce the overall cost of maintaining real-time
balance between generation and load in systems flexibly should be devised and deployed.
with high penetrations of VERs. The design of these mechanisms is an
important area for research today.
Merging or expanding cooperation between
neighboring balancing areas, as well as rein-
forcing when needed the interconnection Finally, beyond the variability and uncertainty
capacity among them, also would reduce the of their outputs, VERs also exhibit other
impact of growing VER penetrations on physical and electrical characteristics that are
operating reserves. very different from conventional generators.
As anticipated VER penetrations increase, it
R E CO M M E N D AT I O N will become increasingly important to design
Fully integrating wind forecasts into VER grid interconnection standards that allow
VERs to enhance grid functionality while
system operations, moving scheduling
remaining compatible with their unique
decisions closer to real time, and expanding
characteristics. The U.S. has made considerable
cooperation among neighboring balancing progress in this area in recent years, and a
areas are all operational changes that variety of organizations are continuing to
should be considered in regions that expect closely monitoring the adequacy of existing
high VER penetrations. These changes can interconnection requirements.
reduce the negative impact of high VER
penetrations on system operating reserve
requirements.

High penetrations of VERs will amplify the
importance of power system flexibility. Flexible
resources will be needed to accommodate the
variation in the output of VERs. Many regions
already have sufficient flexibility in their
existing generation fleets to accommodate the
VER penetrations anticipated in the immediate
future. However, additional flexibility may be

72 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

no.” paper presented at the MIT October 18–19. http://www. 16 North American Electric Reliability Corporation. NY. http://web. Xcel the Federal Energy Regulatory Commission.S. Kirby. Growing Wind: Final Report of the NYISO 2010 gov/corporate/pubs/final-report-columbia-river- Wind Generation Study (Rensselaer. 6 Ibid. U. 2010. 2010).” IEEE Transactions Systems with Large Amounts of Wind Power: Final on Power Systems 25. System (Berkeley. Western Wind and Solar Integration Solutions (Brussels. of Energy. and M. CO: National Renewable Energy American Electric Reliability Corporation. CA. Ela and B. see note 3 above.. RM10-11-000 (2010). Department E. 7 Federal Trade Commission. Notice of Inquiry Re Integration of Variable Energy edu/mitei/intermittent-renewables/papers/ Resources Under RM10-11 (Washington. Department of Energy. and North Study (Golden. Milligan et al. TN. 12 Federal Electricity Regulatory Commission.wind-integration. Generating Assets. Denny. and Eurelectric. Cambridge. Chapter 3: Integration of Variable Energy Resources 73 . April 20.S.. Holttinen et al. Proposed Rulemaking Re Integration of Variable Understanding Variability and Uncertainty of Energy Resources Under RM10-11 (Washington. Photovoltaics for Integration with the Electric Power 2010). presentation at the 9th Annual International Workshop on Large-Scale Integration of Wind 4 S. 14 Eurelectric. 2009).gov/docs/fy08osti/43373. 2011). Laboratory.. Holtinnen et al. 2008: Lessons Learned (Golden. 2009). 17 International Energy Agency.1 Report: Variable Generation Plan_-_June_16. Winter. Wind Integration Study—Final Report (Knoxville. 8 Bonneville Power Administration. Finland: VTT.gov/docs-filing/elibrary. http://www. Design and Operation of Power Significant Wind Penetration. and Bonneville Integration of Renewable Resources (Folsom. Belgium.pdf. 2 (2011): 179–192. CoalImpacts. “Comment of Federal Trade Commission under RM10-11. Energy Outlook (Washington. see note 10 above. Troy.pdf. 2010). CA: Lawrence Berkeley National Laboratory. Hesler.gov/corporate/windpower/docs/WIT_Work_ NERC IVGTF Task 2. Wind Integration Team Work Plan (Portland. Power Administration. E. 2010). and N. see note 2 above. Mills et al. MA. Results of IEA Collaboration.pdf. NY. “Impacts of Large Amounts of Energy. Harnessing Variable Renewables: A Guide to the Balancing Challenge (Brussels. 2 (2010): 1088–1097. ed.bpa.” Wind (Schenectady. Analysis of Wind Generation Impact on Wind Power on Design and Operation of Power ERCOT Ancillary Services Requirements Systems. Docket Energy and the Minnesota Department of Commerce No. DC. 2009). 2008).mit. 2010). Quebec. Belgium. research note 2493 5 Federal Energy Regulatory Commission. 2010). CO: National Renewable Energy Laboratory.nrel. SPP WITF Wind Integration Study (Boston. “Impact of Cycling on Coal-Fired Power Power into Power Systems. 2008). OR. DC: U. 2 California Independent System Operator. Energy Initiative Symposium on Managing Large Scale Integration of Intermittent Renewables. http://www. wwsis.gov/wind/systemsintegration/ (Princeton. MA. ERCOT Event on February 26..asp. http://www.S. Canada.eu/downloads/ Power Integration: An International Comparison.” submitted to 15 EnerNex Corporation and Wind Logics Inc. Report.” library/EWIS_Final_Report. Annual Independent System Operator. DC. bpa. 11 M. New York Independent System Operator. 2010). Operations (Portland. Notice of (Espoo. O’Malley. 2010). Columbia River High-Water 2007). 2010). Charles River Associates. 2011).pdf. “Base-Load Cycling on a System With 13 H.REFERENCES GE Energy and National Renewable Energy 9 Laboratory. and W.ferc. no. NJ.pdf. Phase One 2006–2008. Belgium.html. high-water-operations. U. NJ. European Wind Integration Study (EWIS) (Brussels.S. Energy 14.. Power Forecasting for Operations (Princeton. and GE 10 H. Integrating Intermittent Renewables Sources into the 3 GE Energy and National Renewable Energy EU Electricity System by 2020: Challenges and Laboratory. OR. nrel.. California 1 U. 2011. 2009). Department of Energy. 2004). Accommodating High Levels of Variable Generation http://www. “Operating Reserves and Wind http://www. and A. Energy Information Administration.

see note 3 above. U. North American Electric Reliability Corporation. Team Lead. 29 International Energy Agency. W. and Tools (Princeton. G. 2010). note 5 above. NERC IVGTF Task 1. NJ. GE Energy. Boyle (Oxford. M. see note 10 above. “Comments on RM05-25-000 (2006). 2010). 31. May 11. 2010).” submitted to the Federal Research Workshop—NCAR. Implications of Wide-Area Geographic Observations on Docket RM10-11. (Princeton. Division. New England. How ISOs and RTOs Are Helping Meet This Public Ontario Wind Integration Study. in Transmission Services. NERC IVGTF Task 2. Benjamin.S. CA: Lawrence Berkeley National Docket No.” submitted to the Federal Energy Regulatory Commission. RM10-11-000 (2010). 30 American Wind Energy Association. NY. Energy Regulatory Commission. (Berkeley. report prepared Policy Objective (2007). and J. 2011.4 Report: Flexibility Require- and GE Energy and National Renewable Energy ments and Metrics for Variable Generation Laboratory. NJ. and New York Independent System Operator.. conversation with authors. 41 Procedures. 2007). Increasing Renewable Resources: Final Report (Schenectady. submitted to the Federal Energy Regulatory 28 North American Electric Reliability Corporation. 2010). 21 Federal Energy Regulatory Commission. and A. 39 see note 10 above. 2007).. Mills 36 as a Mandatory Recipient of VER Atmospheric and R. ed. New England Wind Integration Study 31 ISO/RTO Council. see note 13 above. Boundary Layer GE Energy and National Renewable Energy 37 Processes and Applications. and North American Electric Reliability Corporation. Ferreira et al. D. 2010. see note 17 above. Department of Energy. Chandley and W. 74 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . “A Path to 32 and Data Assimilation to Improve Guidance for Preventing Undue Discrimination and Preference Energy. and H. 2006). and J.” submitted to the Federal Energy Regulatory GE Energy and Independent System Operator 43 Commission. see note 2 above. 34 above. “Variable Renewables and the Grid: 35 22 Wilczak. 25 North American Electric Reliability Corporation. Laughton. see note 5 Holtinnen. RM10-11-000 (2010). NY. see note 28 above. Docket Undue Discrimination and Preference in No. Sørensen. Chandley and W. G.” in Renewable Energy and the Grid: 23 National Oceanic and Atmospheric The Challenges of Variability. see note 16 above.4 Report: Operating Practices. Physical Sciences Laboratory.” submitted Diversity for Short-Term Variability of Solar Power to the Federal Energy Regulatory Commission. Commission. Forecast 38 September 7. “Reply Comments on Preventing Federal Energy Regulatory Commission. see note 3 above. Docket No. Hogan.” submitted to the W. RM05-25-000 (2006). 18 S. D. Holttinen. see note 13 above. RM10-11-000 (2011). see note 17 above. Administration. Electric Institute: FERC Notice of Inquiry on the Integration of Variable Energy Resources. Docket Number RM10-11-000. Hogan. 26 C. Notice of Proposed Rulemaking to Include NOAA Holtinnen et al. Federal Energy Regulatory Commission. Docket No. see 33 20 Ibid. National Oceanic and Atmospheric Administration.. A Survey on Wind Power Ramp Forecasting (Argonne. RM10-11-000 (2010). Department of Energy. “Comments of the American Wind Energy Association on VER Holttinen et al. see note 18 above. Wiser. Variable Energy Resources. U. IL: Argonne National Edison Electric Institute.” 27 Ibid. “NOAA Rapid Updated Modeling J. North American Electric Reliability Corporation. Laboratory. International Energy Agency. P. Independent Electricity System Operator and Canadian Wind Energy Association (Schenectady. for Ontario Power Authority. “Comment to Change Intent of UK: Earthscan. Transmission Services. 42 NOI.S. Colorado. Docket No. 19 ISO/RTO Council. Boulder. 2010). Giebel. Docket No. “Comments of the Edison 40 Laboratory. Error of Aggregated Wind Power (TradeWind. Wilczak. System Operations and Wholesale Markets (2011). An Overview.” presentation at the RE Prediction. 24 James Wilczak.

U.K. Department of Energy and Climate Change,
44
U.S. Energy Information Administration, see
58

“Planning Our Electric Future: A White Paper for note 1 above.
Secure, Affordable and Low-Carbon Electricity”
J. P. Deane, B. P. O Gallachoir, and E. J. McKeogh,
59
(London, 2011); Eurelectric, see note 11 above; and
“Techno-Economic Review of Existing and
Poyry Energy, Impact of Intermittency: How Wind
New Pumped Hydro Energy Storage Plant,”
Variability Could Change the Shape of the British
Renewable and Sustainable Energy Reviews 14
and Irish Electricity Markets (Oxford, UK, 2009),
(2010): 1293–1302.
http://www.uwig.org/ImpactofIntermittency.pdf.
Federal Energy Regulatory Commission Staff,
60
International Energy Agency, see note 17 above.
45
“Issued Preliminary Permits for Pumped Storage
U.S. Energy Information Administration,
46
Projects” (Washington, DC, 2011), http://www.
Electric Power Annual 2009 (Washington, DC: ferc.gov/industries/hydropower/gen-info/licensing/
U.S. Department of Energy, 2011). pump-storage/issued-permits.pdf.
L. Pouret, N. Buttery, and W. Nuttall, “Is Nuclear
47
“New York State Electric & Gas: Advanced
61

Power Inflexible?” Nuclear Future 5, no. 6 (2009): Compressed Air Energy Storage,” SmartGrid.gov,
333–341; and M. Debes, “EDF Nuclear Sustainable http://sgdev.nrel.gov/project/new_york_state_
Production: Challenges and Adaptation to Market electric_gas_advanced_compressed_air_energy_
Needs,” presentation at MIT-IEEJ Energy and storage; and “Pacific Gas & Electric Company:
Global Change Workshop, Tokyo, Japan, Advanced Underground Compressed Air Energy
September 30–October 1, 2010. Storage,” SmartGrid.gov, http://sgdev.nrel.gov/
project/pacific_gas_electric_company_advanced_
D. Lew et al., “How Does Wind Affect Coal?
48
underground_compressed_air_energy_storage.
Cycling, Emissions, and Costs,” presentation at
WindPower 2011, Anaheim, CA, May 25, 2011. Denholm et al., see note 57 above.
62

International Energy Agency, see note 17 above.
49
D. Rastler, “New Demand for Energy Storage,”
63

Edison Electric Institute’s Electric Perspectives 33,
Bonneville Power Administration, see note 9
50
no. 5 (2008): 30.
above.
ISO/RTO Council, see note 19 above; North
64
Electricity Advisory Committee, Keeping the Lights
51
American Electric Reliability Corporation, see
on in a New World (Washington, DC: U.S. note 10 above; Holtinnen et al., see note 13 above;
Department of Energy, 2009). and G. C. Tarnowski, P. C. Kjaer, S. Dalsgaard, and
U.S. Energy Information Administration, see
52 A. Nyborg, “Regulation and Frequency Response
note 46 above. Service Capability of Modern Wind Power Plants,”
paper presented at the IEEE Power and Energy
D. Hall et al., Feasibility Assessment of the Water
53
Society General Meeting, Minneapolis, MN,
Energy Resources of the United States for New Low July 25–29, 2010.
Power and Small Hydro Classes of Hydroelectric
Plants (Idaho Falls, ID: Idaho National Laboratory, Lew et al., see note 48 above; and EnerNex
65

U.S. Department of Energy, 2006); and B. T. Smith, Corporation, “Wind Integration Study for Public
“U.S. Hydropower Fleet and Resource Assess- Service of Colorado, Addendum: Detailed Analysis
ments,” presentation at the National Hydropower of 20% Wind Penetration,” prepared for Xcel
Association Annual Conference, Washington, DC, Energy (Knoxville, TN, 2008), http://www.uwig.
April 5, 2011. org/CRPWindIntegrationStudy.pdf.
Electric Advisory Committee, see note 51 above.
54 I. M. Dudurych, “Statistical Analysis of Frequency
66

Response of Island Power System under Increasing
North American Electric Reliability Corporation,
55
Wind Penetration,” paper presented at the IEEE
see note 28 above; and Ela and Kirby, see note 9 Power and Energy Society General Meeting,
above. Minneapolis, MN, July 25–29, 2010; R. Doherty
GE Energy and National Renewable Energy
56 et al., “An Assessment of the Impact of Wind
Laboratory, see note 3 above. Generation on System Frequency,” IEEE
Transactions on Power Systems 25, no. 1 (2010):
P. Denholm, E. Ela, B. Kirby, and M. Milligan,
57
452; and L. C. Dangelmaier, “System Frequency
The Role of Energy Storage with Renewable Performance of the Hawaii Electric Light System,”
Electricity Generation (Golden CO: National paper presented at the IEEE PES General Meeting
Renewable Energy Laboratory, U.S. Department 2011, Detroit, MI, July 24–28, 2011.
of Energy, 2010).

Chapter 3: Integration of Variable Energy Resources 75

67
N. W. Miller, K. Clark, and M. Shao, “Frequency California Independent System Operator,
71

Responsive Wind Plant Controls Impacts on Grid Interconnection Standards Review Initiative: Draft
Performance,” paper presented at the IEEE PES Straw Proposal (Folsom, CA, 2010).
General Meeting 2011, Detroit, MI, July 24–28,
D. Brooks and M. Patel, “Overview of NERC
72
2011; and RenewableUK, “RenewableUK Position
Integrating Variable Generation Task Force Efforts
Paper on Inertia,” Version 3.0 (London, 2011),
on High Penetrations of DERs,” paper presented at
http://www.bwea.com/pdf/publications/
the IEEE PES General Meeting 2011, Detroit, MI,
RenewableUK_Inertia_Position_Paper.pdf.
July 24–28, 2011.
68
J. Brisebois and N. Aubut, “Wind Farm Inertia
North American Electric Reliability Corporation,
73
Emulation to Fulfill Hydro-Québec’s Specific
see note 10 above.
Need,” paper presented at the IEEE PES General
Meeting 2011, Detroit, MI, July 24–28, 2011. North American Electric Reliability Corporation,
74

see note 10 above.
69
T. Ackermann, ed., Wind Power in Power Systems
(Chichester, UK: John Wiley & Sons, 2005).
70
Federal Energy Regulatory Commission, FERC
Electric Tariff No. 3— Section III—Market Rule 1—
Standard Market Design (Washington, DC, 2005),
http://www.iso-ne.com/regulatory/tariff/sect_3/.

76 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

Chapter 4: Transmission Expansion
In this chapter, we consider regulatory policy affecting transmission expansion, with particular
focus on the implications for transmission of the integration of large-scale renewable generation.
Public policies aiming to access the best onshore wind and solar resources will require new
transmission lines crossing state boundaries, independent system operator boundaries, and land
managed by federal agencies. This chapter focuses on the transmission planning, business models,
cost allocation, and siting challenges related to this expansion.
Section 4.1 provides background on the drivers and business models for transmission development
in the U.S. A discussion of transmission planning follows. In Section 4.2, we highlight the importance
of interregional and interconnection level transmission planning. We also find that existing data and
planning methods are inadequate to meet the challenge of renewables integration and highlight
transmission planning under uncertainty for complex networks as an important area for research.
Section 4.3 discusses transmission cost allocation, starting with a review of current practices in the
U.S. We find that cost allocation should be intimately linked to transmission planning. We identify a
set of core principles that should be followed as closely as possible to ensure that cost allocation is
not a barrier to efficient and reliable network expansion.
Section 4.4 introduces challenges related to siting new transmission capacity. We find that current
siting procedures are biased against approving interstate transmission projects and are a significant
hurdle to efficient transmission expansion.
Section 4.5 provides our conclusions and recommendations. We first recommend the creation
of permanent processes for conducting planning of interregional transmission projects at the
interconnection level. We also recommend the compilation of detailed and comprehensive data
on the U.S. bulk power system in order to support research on the methods that will be needed for
effective interconnection level planning. We recommend the use of the cost allocation principles
introduced in Section 4.3 as well as the adoption of a hierarchical approach to cost allocation
that includes a single, uniform procedure for the allocation of costs between regions within
each interconnection but allows individual regions to adopt their own internal cost-allocation
procedures. Finally, we recommend that Congress grant the Federal Energy Regulatory Commission
enhanced siting authority for interstate electricity transmission projects.

Several factors beyond normal growth in renewable generation, such as large wind farms,
electricity demand will require new investment that will be connected to the high-voltage
in transmission capacity in the next two transmission system.
decades. Perhaps the most important is the
need to integrate large-scale renewable genera- Also driving expansion, the recession-induced
tion. The U.S. federal government and all state fall in electricity demand and the shale gas
governments provide financial support for the revolution have lowered electricity prices and
use of renewable energy to generate electricity, adversely affected the economics of operating
and 29 states and the District of Columbia have many older coal-fired units. A suite of new
enacted quantitative requirements.1 These rules from the U.S. Environmental Protection
programs will spur growth in large-scale Agency may provide further incentives for early

Chapter 4: Transmission Expansion 77

retirement of those units. A number of recent Some have argued that the best solution to this
studies have concluded that, as a result, signifi- suite of problems is to construct a nationwide
cant early retirements are likely in the coming overlay or “super-highway” grid.5 Others favor
decade.2 Early retirements and changes in large, discrete transmission projects that
dispatch will change the geographic pattern connect sizeable renewable resources to major
of generation and require new investment in load centers.6 Still others defend a more
transmission facilities. conventional buildup of transmission rein-
forcements—within regions and across
In the reference case of its 2011 Annual Energy multiple regional boundaries—and more
Outlook, for instance, the Energy Information use of local renewable resources.7 We do not
Administration projects that 46% of the recommend or oppose any particular suite
increase in total generating capacity in the of investments. Given the complexity of
electric power sector between 2010 and 2030 transmission expansion and the many
will be powered by non-hydro renewables, competing alternatives, this is a matter for
which accounted for only about 4.5% of 2010 careful decision-making by knowledgeable
capacity. Almost 90% of this increase is stakeholders. Only by unlikely coincidence will
projected to come from growth in wind and the public interest be served by transmission
solar capacity.3 plans dictated by legislation or based on
comprehensive visions devised from afar.
Wind and solar generators have two character- Our focus is on the planning, cost allocation,
istics that challenge transmission systems. and siting processes and criteria by which
First, even when fully functional, their available important transmission investment decisions
output can vary from zero to full capacity in should be made.
relatively short times and is less predictable
than output from other generation tech- Consistent with this focus, the Federal Energy
nologies. We discussed the implications in Regulatory Commission’s (FERC) July 2011
Chapter 3. The second characteristic, as noted Order No. 1000 required improved coordina-
in Chapter 1, is that many of the best onshore tion in transmission planning and cost alloca-
wind and solar resources are located far from tion procedures within planning regions and
major load centers and are, therefore, far from between neighboring regions.8 (The Order does
the existing transmission system. If they are not define “region,” but notes that planning
to be tapped efficiently, an increasing fraction regions are smaller than the two larger inter-
of transmission lines will cross state borders, connections and larger than single utilities.) We
independent system operator (ISO) regions, believe the Order is a step in the right direction
and land managed by federal agencies such as but that the public interest would be best
the U.S. Forest Service. While some boundary- served if affected parties went beyond the
crossing lines have been built in the past, that order’s minimal requirements. Moreover,
experience underscores a number of obstacles FERC’s limited authority prevented it from
such projects face.4 As we discuss in Section 4.2, addressing the problem of siting transmission
there is too little useful transmission planning facilities that cross state boundaries or federal
at regional, interregional, or interconnection- lands. We argue that this process, too, needs
wide levels; the costs of boundary-crossing reform.
projects can be allocated only through project-
specific negotiations; and the need to obtain
construction permits from multiple authorities
makes it difficult to site and build boundary-
crossing lines.

78 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

4.1 TRANSMISSION DEVELOPMENT as a fourth category, and they are explicitly
IN THE U.S. recognized in FERC Order No. 1000. While
these labels are common, in practice, increased
Between 1980 and the late 1990s, annual transmission capacity will provide the system
investment in the U.S. transmission system with multiple benefits that may change over
declined in real terms, and many observers time. While it has been convenient to label
expressed concern that the system was a transmission line according to its primary
becoming economically inefficient and purpose, this convention is at odds with the
unreliable.9 A careful analysis of available new reality and could be both confusing and
data suggests that this concern was mostly counterproductive. Any transmission line
unwarranted.10 Transmission and distribution serves all these purposes to different degrees.
losses generally declined over this period In the future, coherent policy for analysis of
(see Figure 1.4), and there is no quantitative costs and benefits should recognize and capture
support for assertions of diminishing reliability. this interaction. Aggregation across the benefit
Moreover, since the late 1990s, investment in categories should be the norm in evaluating
transmission has increased considerably, and transmission lines.
these earlier concerns are less often heard today.11
FINDING
The transmission system is not broken, and Transmission lines routinely serve a
there has been and continues to be substantial
variety of purposes with many categories
investment in system upgrades and new
of benefits. Aggregation across benefit
interconnections. Today’s grid meets today’s
requirements, but new and different demands categories should be the norm in
are driving the expansion and adaptation of evaluating transmission lines.
the transmission grid and the evolution of its
supporting institutions. The system can and
will respond to the new forces, but effective Reliability is the most common justification
response will require material changes in the for transmission investment in the U.S.
regulatory and policy framework. Transmission projects are developed either
to meet reliability standards promulgated
Drivers of Expansion by the North American Electric Reliability
Corporation (NERC) and regional reliability
U.S. transmission projects traditionally have authorities or to accommodate uncertain future
been categorized by the primary purpose growth and development without violating
they serve: reliability, economic efficiency, those standards.ii, 12 The benefits of reliability
or generator interconnection.i Public policy are difficult to quantify and are often asserted
purposes—for example, to meet renewable to be spread over relatively wide areas.
generation targets—have recently emerged

i
As a general matter, it is important to recognize that non-transmission investments can sometimes serve the
same purpose as transmission investments and that transmission investments do not always take the form
of new towers and wires on new rights-of-way.13
ii
A discussion of the reliability standards that are used to govern generation and transmission investment
decisions is beyond the scope of this study. For the discussion in this chapter, we take them as given.
However, the criteria underlying current reliability standards do not necessarily reflect rigorous cost–benefit
analyses, an issue FERC is addressing.14

Chapter 4: Transmission Expansion 79

The frameworks for assessing the economic value of these reliability upgrades tend to be either very weak or nonexistent in most regional transmission organizations. 80 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . solar and wind plants at locations far from the existing network. there is growing concern that current trans- In addition. Electric Reliability Council of Texas (ERCOT). California. and Southwest Power Pool (SPP) have planning iii This point is well illustrated by Midwest ISO’s 2009 transmission expansion plan. clear technical procedures for do planners look for investments that would justifying reliability investments have generally increase economic efficiency. while lines justified or eliminated capacity constraints (generally primarily by economic benefits are rare. narrowly defined. such lines have been short Economic planning processes are generally not nearly and largely uncontroversial. Historically. if companies develop large completing their first studies of economic opportunities. and many regions are still in the established practices that account for policy process of completing their first studies of objectives. and they have been as well developed as the procedures for reliability included in proposals for new generating planning. is difficult to make in the current recession.18 termed “congestion”) that prevent the use of Yet we cannot identify specific situations in the lowest-cost set of generators to meet which lines with clear economic benefits were demand. which included $4 billion worth of exclusively reliability lines that were expected to provide nearly $3. current lines built that had been justified primarily on treatment of these lines and required reinforce- the basis of economic benefits.4 billion in economic benefits. though. as renewables receive increasing a persuasive economic case for investment amounts of attention in public policy debates.1). and many regions are still in the process of capacity. most available economic benefits are captured by lines built primarily for reliability. and Minnesota have already planning. perhaps network. economic opportunities. Of course. while the most basic economic mission development procedures may not benefits are in principle easier to quantify than adequately support their development.19 which mitigates market power and may provide other benefits. lines justified by Generator interconnection lines allow genera- economic benefits generally improve system tors to connect to the most appropriate point reliability and vice versa.17 Economic tion has emerged in some regions and has been planning processes are generally not nearly as endorsed in FERC Order No. Reflecting this concern. the concept of measuring them in practice is a challenging including public policy purposes as a justifica- analytical task (see Box 4. the equivalent generator interconnection lines may be longer and more Recent years have seen very few transmission expensive than in the past.16 The stringency ments elsewhere in the network may need of reliability requirements may ensure that revision. By strengthening the transmission planned but could not be built.iii. and California ISO. Midwest ISO. Economic benefits ensured that all the lines necessitated by include reduced network losses and mitigated reliability are built. 1000. However. well developed as the procedures for reliability Texas. Colorado. 15 on the transmission system—usually the closest. the economic value of reliability benefits. and Indeed.20 Midwest ISO’s systematic assessment of these investment benefits is unusual. As a result. Only after reliability planning is complete In the end. these projects also allow wholesale because economic benefits have been too electricity markets to expand geographically.

In planning. Some projects to deliver renewable power will FERC Order No. Unless project proponents the environment. subsequent investments and assumptions sion congestion isolating them is alleviated. Neither development. but jobs. one-size-fits-all solution. and mission expansion. However. The separate but Benefits assessment is complicated in practice related problem of appropriate cost allocation by the fact that transmission affects not only also presents significant challenges for trans. where improve the economic efficiency of the power planning criteria are examined. and the costs and system over the short and long term. local interests. such as the economic value problem has a simple. system requires a calculation of the economic benefit of that investment. symmetrically. how to value some a project forward in the face of inevitable hard-to-quantify outcomes of transmission opposition in the siting process. a complica- change in the least-cost way of reliably meeting tion that is made all the more problematic by that demand that the project would make the long lives of the assets involved. At the most basic level. it will be difficult to move benefits to consider. regarding what would have been built instead Benefits are also. generation resource additions and the project. unlike most residential because defining the counterfactual would and other small customers. Note that of lines has not been built. BOX 4. nor will they necessarily and regional transmission planning processes bring economic benefits until state or federal include policy objectives. Chapter 4: Transmission Expansion 81 . Planners benefits have been fairly apportioned among and stakeholders need to determine which the affected parties. solution. technological change. generators in high-price areas when a new line processes that allow for public policy benefits. and what approach but as we discuss in subsequent sections. It also mandates that policies sufficiently reduce the cost of renew- public utility transmission providers establish ables or increase the cost of fossil-powered procedures to identify transmission needs generators. electricity prices. transmission investments. of reliability improvements. This is not practical these benefits will be negative for consumers with lines that were built a long time ago in low-price areas who. Allocation of project costs should be their benefits.1 THE CHALLENGE OF ASSESSING increases the linkage with areas with lower ECONOMIC BENEFIT electricity prices. To assess possible. the economic benefit Beyond the particulars of quantifying benefits. 1000 will require that all local not increase reliability. large transmission can persuasively argue that the benefits are projects often add important flexibility to the likely to outweigh the costs. and relative forecasting future demand and finding the fuel costs are grounds for debate. In addition. of any project is measured by the increase in assumptions about future environmental consumer surplus plus producer profits from policies. new lines can be compared to a based on benefits to market participants— counterfactual situation in which the line or set both consumers and generators. this translates into retirements. negative for of the line in question. A more comprehensive view of transmission-related benefits is given in the Justifying investment in transmission to “Planning Criteria” section on page 87. will be taken to determine the “optimal” reasonable and workable solutions do exist. the policy dimension can driven by public policy requirements and be made commensurable with economic and evaluate alternative ways of meeting those reliability benefits for purposes of evaluating needs. face location-specific require the hypothetical untangling of all the rather than area-average prices when transmis.

and how its costs will be recovered. 1000. The idea behind merchant investments is to sion. Hawaii. the Alternatively. but it is beyond the scope of our study. which lowers transmission expansion as part of an integrated barriers to their participation. v The corresponding contract rates may be similar to those that result from standard regulatory processes. build. Tariff-financed projects undertaken by non-incumbents are also rare In regions without an ISO.S. and other entities not subject to FERC regulation.22 resource planning process. about one-third of the high-voltage transmission system in the 48 contiguous states is owned by government enterprises. 82 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . Once the plan is approved by the relevant state utility commis. but this may change as the FERC integrated utilities will each centrally plan implements Order No. In others. in which costs are electricity transmission is generally assumed typically recovered through contracts between by law to be in interstate commerce and thus the transmission owner and specific users who subject to FERC regulation because it takes benefit from the investment. In most cases.v rates. investors may try to cover their division of costs between them is negotiated costs by arbitraging the differences in the and. The legal and policy problem of harmonizing their behavior with that of the entities regulated by FERC is complex and important. appropriate for voluntary or “merchant” transmission investment. however. investments are lumpy. although they also may be higher because merchant investors are not prevented from seeking a high rate of return when bargaining with beneficiaries. Transmission Investment Regions with an ISO have more latitude for different investment schemes because multiple During and following the processes of trans. the local vertically in the U. utilities and ISOs who will invest in and build the new line and identify needed network improvements. The project costs are allocated to increase reliability. non-incumbents may propose Some projects to deliver renewable power will not projects. after approval by the state regulator or locational marginal prices between the two regulators. In addition. In other cases. the utility builds and maintains the lines create new transmission capacity where signifi- that have been proposed and recovers the total cant locational price differences or important cost—including an allowed rate of return on network constraints exist that the new line the investment—through state-regulated retail will reduce or eliminate once it is in service. nor will they necessarily bring network users based on a wholesale transmis- sion tariff. respond to reliability iv Alaska. the transmission level in coordination with the states’ public builder is always an incumbent—a transmis- utility commissions. But because transmission through its own retail rates. and the corresponding sufficiently reduce the cost of renewables or increase retail tariffs. FERC plays a role as well: sion company that already has a presence in the region. Voluntary funding place in an interconnected high-voltage grid of large-scale transmission projects is that crosses state lines. parties may own. each recovers its share of costs ends of the line. In some jurisdictions. proposed by the ISO or utility and economic benefits until state or federal policies approved by FERC.. cooperatives. When lines connect two utilities. the conditions are the cost of fossil-powered generators. and operate transmis- mission planning. These processes a transmission utility builds and maintains the take place primarily at the state or regional project. it is necessary to determine sion assets.21 These entities have much higher shares in some regions.iv uncommon. and the ERCOT region of Texas are exceptions.

making significantly more difficult than within vertically it difficult for investors to cover their costs in integrated utilities because decisions about the this fashion. Because NERC regional entities planning. mission planning in these regions is subject sored by new entrants and generally involve to additional uncertainties about where future high-voltage direct current (dc) technology generation may locate and how power will flow (discussed in Chapter 2). however. load is served by large. which considered wind ning processes. indi.25 ISOs. but they focus primarily on regional projects but had not developed formal the objectives of maintaining a transmission interregional planning processes.26 Magnifying this effect of the line’s benefits through their ability to are uncertainties regarding future subsidies and control power flows. especially when renewable owners of the facility to capture a larger portion generators are involved. The U. process and thereby potentially discourage new and interconnection accelerated thereafter. and build a high-voltage transmis- sion line than a wind farm or solar generating As noted in Appendix A. problems could arise that would (see Figure 1. When generator build times are shorter the industry.24 a painful fact of transmission planning is that it typically takes much longer to plan. and vertically integrated ering interregional projects one at a time rather utilities do little joint planning. generation investment. they tend to The planning process in most ISO regions is reduce price differences substantially. But the grid that is reliable and.23 Would-be merchant investors installation of new generation are the result may find it difficult to reach an agreement with of market forces (modified by state and federal enough beneficiaries willing to help cover the support for renewables and other policies) cost of the line. which allows the around the network. Transmission Study. trans- exist or have been proposed are mostly spon. regional ISOs that plan Interregional Transmission Planning transmission to serve projected loads in their territories. though this than as parts of an interconnection-wide plan is changing as interconnection increases is no longer sensible.S. Some interconnection linking these for firm generation plans before starting the utilities occurred before World War I. transmission planning was the than those for transmission. penetrations of 20%–30% in the Eastern Chapter 4: Transmission Expansion 83 . The planning processes vary Until recently. Indeed. about two-thirds of the U. power marketing becoming more tightly coupled. in the early years of facility.3) are responsible for reliability impede efficient expansion of renewables throughout their geographic domains.2 TRANSMISSION PLANNING approvals. Department of Energy vidual vertically integrated utilities within the (DOE) Eastern Wind Integration and same NERC region tend to have similar plan. and have scale economies.S. ISO regions and states had significantly among different ISOs and worked together to negotiate specific inter- non-ISO regions. with lower priority. Thus. because to define the beneficiaries. which makes it easier requirements for renewable generation. get 4. Eastern and Western Interconnections are economically efficient. planners are forced responsibility of vertically integrated utilities to either anticipate new generation and build that met their native loads from their own potentially unnecessary infrastructure or wait generation. Today. generation. without such over time.criteria. The few merchant projects that rather than centralized planning. and consid- administrations.

27 for the Western Interconnection. and top-down processes.Considering interregional projects one at a time The Western Interconnection has long been rather than as parts of an interconnection-wide plan a leader in wide-area transmission planning using a hierarchical approach.29 shortcomings: A solely bottom-up approach In addition. Transmission System Operators for Electricity chical hybrid of the two approaches has the (ENTSO-E) has been tasked with providing a potential to respect local and regional needs 10-year pan-European transmission expansion while still having vision broad enough to plan and has recently finished its first prelimi- recognize interregional opportunities. nary report. members of the Western Interconnection. But a purely top-down process may not be adequately responsive to By comparison. the study concludes that reaching very to model economic transmission expansion high penetration of renewables will require through the Transmission Expansion Planning substantial use of local and remote wind Policy Committee (TEPPC). one needs in effect since 2004. the European Network of regional issues or planning processes. Interconnection. expansion of high-voltage lines. which spans the resources.28 While TEPCC and its panying transmission development within and subcommittees model the strategic economic across multiple regions. planning exercises. the scenarios prioritizing local wind. In the Eastern Interconnection. the Northeastern ISO/RTO will fail to identify potentially desirable lines Planning Coordination Protocol among ISO that traverse regional boundaries. the Public Service Company of Colorado participates in TEPPC. until recently. New York ISO. Both have their top-down regional planning processes. demonstrates that even in Electricity Coordinating Council (WECC). In the Western is no longer sensible.and intraregional lines. transmission plans that are based on detailed knowledge of local or regional conditions. the council also has recently interconnection-wide planning requires a started an electric transmission planning study hierarchical approach encompassing bottom-up for the entire Western Interconnection. and perhaps interconnection. A hierar. and state planning procedures. the East top-down processes. To capture New England. the regional transmission planning group WestConnect. collaborate Further. 84 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .vi With American Recovery and Reinvest- and Western Interconnections are such that ment Act funds. which in turn will require accom. to plan transmission for the state of Colorado. both the PJM Top-down planning involves a central body Interconnection and SPP have bottom-up charged with identifying potentially desirable subregional planning processes to supplement inter. tion comparable to the TEPCC in the West. entire interconnection. smaller subgroups model reliability and lower-voltage The scale and complexity of the Eastern lines. Bottom-up planning although this research activity may or may is the process of integrating local or regional not influence what actually gets built. The newly created Agency for the Cooperation of vi For example. and PJM has been these potential investments. wide. but a directive establishes that national plans should be consistent with the pan-European one. performed as part of has not had an interconnection-wide institu- interregional. transmission NERC regional reliability council responsible requirements span multiple operating regions.30 This plan is not mandatory. But. the Colorado Coordinated Planning Group within WestConnect.

as well as an and one in ERCOT. interconnection level. to its jurisdiction to participate in these interregional cant deviation to the European Commission. Moreover. two top-down analysis necessary to complement each in the Eastern and Western Interconnections existing bottom-up processes. In ness and impact on what actually gets built addition. The West assistance in the form of “modeling. planning processes. of authority or bureaucracy.31 No. We discuss these next. but it is noted that FINDING a single utility cannot constitute a region for In the Eastern and Western Interconnec. FERC urges which has become more important. FERC’s Order to 2050. though it cannot require them to do so. interconnection-wide planning.32 “Region” is undefined in the order. however. Two recent developments may serve to expand While existing planning arrangements have the effective scope of hierarchical wide-area enabled construction and ongoing expansion planning in the U. because of data availability will need to be addressed. Recently ENTSO-E has started to develop a planning processes. problems of planning methods and remain to be determined. 1000 requires regional and interregional planning between adjacent regions. but their effective. requires transmission facility owners not subject to its a hierarchical approach encompassing both jurisdiction to participate in these interregional bottom-up and top-down processes.Energy Regulators is responsible for supervising FERC urges transmission facility owners not subject this consistency and for reporting any signifi. particularly in the East. vision for how a pan-European power system could develop sequentially over a time horizon In a second development. across each interconnection. Act appropriated $80 million to establish extensive use of renewable resources distant interconnection-wide planning collaboratives from major load centers will require stronger. Chapter 4: Transmission Expansion 85 . with the goal of “facilitating the development permanent interconnection-wide planning of regional transmission plans” and providing procedures. though some laws. the new collabora. though it cannot require them longer-term strategic plan that will provide a to do so. and adequate stakeholder participation.” The Office of Electricity central staff and modeling capability will Delivery and Energy Reliability at the DOE presumably be required to perform the issued awards to five separate organizations. programs. and regulations. support has demonstrated that this can be accomplished to regions and States for the development of collaboratively. tives are only supported for a single round of analysis and may well disappear afterward. this purpose.S. a made up of regional planning authorities from prescribed periodicity of the plan. Title IV of the Recovery of a reliable and efficient transmission grid. the nature of their funding. These organizations are agreed-upon definition of planning criteria. It may not require another layer coordinated State electricity policies. transparency. There are no requirements at the tions.

system rein- Making more use of remote renewables forcements or other remedies are developed. many possible reinforcement options can resolve system concerns. geographic areas and with the increasing levels of uncertainty that must be considered To evaluate a network design’s robustness. models. Some optimization techniques are dimensions. and locations.34 magnified when planning over larger areas and trying to achieve multiple objectives. in an efficient manner will require Next. transmission planning is characterized a time. These character- used in some systems for a long time. environmental policies. large in principle capable of producing system plans investments. fuel prices. experts frequently define long-lived modifications to complex networks the set of possible reinforcements.33 However.35 Moreover. and they have been investments must be assessed. uncertainty. the transmission system is planned during short time frames but may enable the using expert judgment supported by technical efficient evolution of the grid in the long term. simulations identify reliability issues and potential economic improvements. Because the transmission system is a complex Transmission Planning Methods network. the simulations are re-run to ensure that permanent planning processes at the the reinforced system meets the prescribed reliability requirements and delivered energy interconnection level. As noted above. Transmission planning is characterized by a large the impact of uncertainty surrounding plant number of choices with multiple dimensions. including moderate levels of uncertainty of transmission investments. although istics are compounded and the challenges they have limitations of their own. and generation portfolios may is able to address power systems on an ISO vary substantially over the 50-year lifetime level. because load characteristics periods over which investments must be assessed. a great location is often compounded by the mismatch deal of uncertainty. and long between generation and transmission build times. FINDING simulations indicate a problem. which allow them to consider investments that may not be deemed prudent Today. Complex a computational and conceptual challenge. In contemporary Transmission planning involves discrete and transmission planning. The Restructuring and the ensuing separation current state of the art in transmission planning of transmission and generation planning will increase uncertainty. More techni- planners tend to consider one investment at cally. to integrate substantial renewable generation planners perform multi-period analyses under efficiently. current methods be designed to perform well under a variety have not dealt with planning over the larger of different conditions. a great deal of uncertainty. Expert in the face of an uncertain future. and long periods over which without this limitation. the network must on a scenario basis. costs are reduced. however. large investments. If the 86 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . rather than focus on system by a large number of choices with multiple outcomes. The general procedure is to forecast Performing such analyses for a complex network demand 5–10 years into the future and simulate subject to multi-dimensional uncertainty is the system performance at that time.

accessed.40 clear criterion in transmission planning FINDING initiatives. regulatory and other uncertainties regarding These states may fail to meet their renewables the availability of renewable resources. including security growth in renewable generation but that concerns. and little work has been done to criteria are often left implicit and may include develop methods to support robust network robustness and flexibility. have been used in some cases. and the proposed line’s policy goals also play a formal role. In cases in which the criteria are unclear. such as the Renewable Energy Transmission Initiative (RETI) in California and the Planning Criteria Competitive Renewable Energy Zone (CREZ) planning in Texas.39 renewables coupled with the mismatch between generation and transmission build times make Finally. explicit criteria for trans. The most obvious conflict today can occur in states with Scenario methods. the environmental requirements for renewable generation. high-cost resources. Forward-looking studies often consider sale markets and mitigation of market power. yield an optimal expansion path to the eventual desired network. ness to situations in which the envisioned conflicts can arise between system planning scenario does not unfold. In states or regions with renewable energy accessed.however. regarding future support and requirements for and then employed. to improve planning methods for such adverse outcomes more likely. stochastic planning criteria. Transmission to scenario processes. methods are not adequate to support interconnection-level planning that takes Some authorities have recognized this problem appropriate account of uncertainty. Uncertainty by the industry and the research community. which consider multiple renewable portfolio standards but no trans- futures. nor do they consider robust. the commercial viability of the economic efficiency. criteria for assessing lines include the quality mission planning include NERC and state and quantity of the renewable energy resource reliability standards and. States or regions with policies driving For a variety of reasons.36 These analyses do not different localities. only the design of networks for a static year and the ease of constructing transmission in and single scenario. may fail to meet their renewables Currently available data and planning requirements for want of transmission. processes and policy requirements. In RETI. detailed data used by transmission operators on real wide-area FINDING networks must be available to use for testing. public impact of a new line. wide-area networks. in many regions. and established explicit transmission directives. Additional ability to bring renewable energy to market Chapter 4: Transmission Expansion 87 . for instance.38 requirements for want of transmission. expansion of whole- planning. or they Because increased uncertainty cannot always may meet them using unnecessarily low- be dealt with adequately via deterministic or quality. serve renewables will not materialize without tools. such data are not generally available do not establish public policy goals as a for this purpose today.37 But mission initiatives beyond the typical general scenario methods may not identify important requirement to provide transmission access. and methods will need to be developed appropriate policy requirements. the As noted earlier.

some generalizations are possible. However. Costs of to the transmission planning process. economic efficiency. and FERC order and then explore their application generator interconnection purposes. The costs of lines through regulated rates over time. Rather than proposing a specific cost-allocation In most ISO regions. nor allocation method for all utility transmission have they instituted consistent cost allocation providers and neighboring regions must have a procedures across or between regions. No.42 common interregional method. This general approach is commonly referred to as “socialization” of An important policy challenge is to develop costs. 1000 mandate that regional transmission the ISO regions have not yet converged on planning processes must designate one cost- common principles of cost allocation. projects justified solely on reliability grounds are usually recovered on a uniform basis from Principles of Cost Allocation electricity users.43 Even so. that cross boundaries between regions are divided between the two regions in a manner As we noted earlier.41 These initiatives may remaining costs are socialized. allocation. In network users through a separate charge. upgrades are likely to be more expensive on most transmission costs are recovered from average than they have been historically. new generator interconnection lines and related system But in ISO regions with wholesale markets. On the other hand. The corresponding cost has severed what had been a clear association allocation rules and processes will be under between specific transmission assets and the scrutiny and modification due to FERC’s Order generators and loads they served. Projects that are necessary to reinforce the network because As discussed earlier. To date. Individual users’ shares of total costs principles and procedures for mandatory cost typically depend on total or peak consumption. addition. Participant funding for various forms of merchant lines or sale of capacity rights on large dc lines provide other cost 88 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . regional cost allocation method. which creates economic grounds are sometimes allocated in transparency and allows stakeholders to see whole or in part to those end users who benefit why particular transmission expansions were most from their construction. interregional lines probably will be This shift from utility-by-utility allocation more important.3 TRANSMISSION COST ALLOCATION not done in a consistent fashion. transmission costs are generally the costs of radial lines connecting generators recovered across the utility’s entire territory to the grid are often refunded to generators through regulated retail rates. even markets. this is 4. The RETI The costs of projects justified primarily on and CREZ criteria are detailed. as opposed to relieving congestion. cost allocation differs a new generator is being connected may be by market type. we next lay out some fundamental is performed differently for projects that serve principles that complement and extend the primarily reliability. if intensive use is made of negotiated specifically for each project. The generator involved tends to pay for gener- ator interconnection projects. be possible models for broader multistate or national initiatives. and any studied and adopted. the best renewable resources. In regions served by vertically allocated to loads and treated like reliability or integrated utilities without organized wholesale economic upgrades.

if its benefits exceed its costs. However. done.S. In addition. this is complicated and seldom.44 This so-called “benefi. if a proposed line would cross a particular state but confer no benefits on its residents. taking negative net benefits overall. benefits for some that exceed costs but with tures or profits as a result of the project. if the project has positive net benefits. which in ¶585(1) states. Costs should be allocated in (see Box 4. and the associated costs are largely fixed ex post. better off and less likely to oppose progress on ties in a manner that is at least roughly the project. Chapter 4: Transmission Expansion 89 . benefits would be willing to cover the costs. ciary-pays” principle has been widely accepted in the U. This means turning into account the value of increased reliability down some projects for which those who receive and any other benefits. but strict adher. these projects can affect the economic value of any existing transmission rights and contracts Principle 1. requirements. And if a project’s allocated to those within the transmission benefits exceed costs. This is the most losses because of environmental harm. it is possible to compensate any losers for their losses and make all affected entities better off. In principle. the scale process has produced a set of estimates of and scope of transmission investments can have expected benefits. generators in previously high-price areas or on ence to them would go beyond the order’s load in previously low-price areas.” We see exceed its benefits.viii Conversely. approving any project with positive net bene- tical to its share of the project’s total benefits. Thus. 1000. In practice. it will be impossible to vii Large. discrete transmission investments can have a material effect on prospective market prices and the distribution of benefits. and abroad. for instance. even if it imposes losses on some entities. It is generally argued that compensation is not deserved for the loss of economic benefits (high prices to generators. it is better to use a multipart charging mechanism that couples ex post efficient usage pricing (reflecting congestion and marginal losses) with fixed access charges. The common feature of no principled reason not to take this one step these alternatives is that they are voluntary. further: the allocation of costs should be exactly Voluntary participant funding is desirable when proportional to those estimates if the planning it is workable. in proportion to their benefits is generally “The cost of transmission facilities must be perceived as equitable. there is a require- ment for regulatory approval and the associated A transmission project is economically justified mandatory rules for allocation of the costs. It stands at the core of Dividing a project’s costs among network users FERC Order No. and some entities might suffer proportion to benefits. By reducing or eliminating price differences. all beneficiaries will be planning region that benefit from those facili. but major environmental impacts may raise more serious issues in the future. This stronger language would a material effect on market conditions and avoid an interpretation permitting cost alloca- create concerns over free-riding that would be tions that depart materially from the pattern of impractical to overcome through strictly estimated benefits. 1000. voluntary agreements. if a project’s costs commensurate with estimated benefits. Such impacts might be claimed.vii fits. Each beneficiary’s share tors can cut through this tangle of effects by of a project’s costs should be as close as prac. Regula- fundamental principle. if ever. viii In principle.allocation models. Rather than recover those costs through ordinary transmission tariffs. in many cases.2). assigned to members of the coalition of putative beneficiaries in a way that preserves net benefits for being a member of the coalition. however. a None of the principles we present here conflicts transmission project could impose losses on with FERC Order No. low prices to loads) that exist only because of network congestion. beneficiaries are any network users They should disapprove projects with gross who see a change in their expected expendi.

the This definition of the marginal cost of transmis. Of course. the economic dispatch with locational prices. as well as determining throughout a region.46 For transmission expansions. which spreads it uniformly what should be built. But the integrity their own trades and determine the final of the FTR does not depend on the actual use of dispatch of the system. simultaneous feasibility rule applied to the sion does not require separating transactions to existing plus incremental FTRs awarded with describe the complex physical flows between the expansion would guarantee the same locations. Another implication Users of the system pay. which spreads it uniformly throughout a region. the set of FTRs operates as problems and support competition by supple.2 FINANCIAL TRANSMISSION RIGHTS charged as the short-run price of transmission. because not enough of the benefits have been captured to cover the costs. BOX 4. and owners of the FTRs is that there is no workable definition of collect. nated and centralized final dispatch built on the For a given transmission grid. the laws Users of the transmission system pay this price of physics dictate that power flows distribute for explicit bilateral transactions or implicit across all possible paths between locations. This is an inherent property between any two locations is equal to the of economic dispatch and the related locational marginal cost of transmission between them. Fairness is locational signals.45 The difference in prices holders of the FTRs. reducing the system’s ability important. in particular— regardless of their location and impact on 90 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . awarded are simultaneously feasible. though there were a set of fully tradable and menting decentralized trading with a coordi. the financial transmission right (FTR) that allows sees generator dispatch and maintains balance the owner to collect the difference in locational in the system while respecting the many prices for a given volume between two points. This is need for a central system operator who over. beneficiaries—the generators. trades through the coordinated dispatch. the system corresponding to the distribution of Organized electricity markets overcome these FTRs. for investments is the key reason for embracing For instance. if all the FTRs framework of bid-based. socialization would this principle. transmission constraints. approximately minimizing total system losses. Thus adopting the benefi. The difference in locational prices also provides This creates strong interactions between a means to define an economic alternative to generators and loads. The revenues collected from the system users will prices reflect the locational marginal values of be sufficient to cover the payments to the generation and load. but support of consistent incentives to promote investment in the best locations. reconfigurable physical transmission rights. The difference in locational prices is property going forward. An inferior but commonly used alternative to the could cause a beneficial project not be built beneficiary-pays principle is the socialization of cost. In effect. to the beneficiary-pays principle is the social- ciary-pays principle helps with decisions about ization of cost. security-constrained. all else equal. allocate costs in such a way as to make all An inferior but commonly used alternative entities better off. In an electricity transmission network. failure to recognize all always favor the best wind or solar resources. prices. The economic effect is the same as “physical” transmission rights that would allow would be true if it were possible to define and for individual generators and loads to arrange use separable physical rights. Socialization eliminates who should pay for what is built. One implication is the the missing physical transmission rights.

ix This danger is illustrated clearly by PJM tariff submissions. network charges are levied as an annual lump cantly exceed the benefits they realize. Great generators operate in highly competitive uncertainty about benefits and beneficiaries environments. if whole- significant changes in decision-making in the sale markets are highly competitive and there electric system and put many important are no special opportunities for any generator investment decisions into the hands of regula. some generators may enjoy a workable approximation when much uncer. same as abandoning the principle. Moreover. this argument misleads. it does employ one commonly employed proxy for benefits.1. alternatives.ix sum or on a per megawatt basis rather than per megawatt-hour of produced energy. not all However. but doing so would call for As in any highly competitive market. The beneficiary-pays generators to enjoy short-term rents (and suffer principle is still applicable. doing so. however. to the generators involved. as shown by the illustrative example in Figure 4. and changing market conditions generally implies that expected benefits are typically provide multiple opportunities to widely distributed. uniform region-wide cost recovery generators will end up being passed on to load can provoke substantial public opposition to via wholesale electricity prices. allocating too much of the transmis- beneficiaries of new transmission capacity. In some It is sometimes argued that cost socialization is markets. which show a divergence as large as $1. nor would it produce the same result in the more common If regulation fails to allocate costs according cases where significant uncertainty about some to benefits generated and. too.transmission costs. Additionally. sion cost to load would eliminate locational Generators use the transmission system to signals to generators. This is true even if parties are forced to shoulder costs that signifi. especially for renewables deliver their product. is built to these locations. tricity.2 billion in cost-allocation outcomes of a specific project for socialization compared to the PJM DFAX flow-based method. benefit financially from that require costly transmission investments. and should therefore be responsible These signals help to ensure that the most for paying for a fraction of the network costs. Cost-allocation procedures eliminate the incentive to consider economic should seek to apportion the costs of a line to alternatives to transmission expansion. so these generators can be might produce a cost allocation similar to charged transmission costs without any antici- direct socialization. generation and load generally are both Likewise. or both. they will retain benefits from transmission that or when the investment impacts several regions. The Florence School of regulation has provided a useful comparative analysis of cost allocation methods.47 One generation and load proportional to aggregate solution might be to socialize the costs of the economic benefits realized by the two groups. unique location-specific or other advantages. for instance. socially beneficial investments in generation could be abandoned. But this would not be the pated pass-through to consumers. economically sensible sites are chosen for Load also benefits from new transmission generator development. Finally. Where there are wholesale markets for elec.49 Chapter 4: Transmission Expansion 91 . spreading through reduced energy costs. even though it short-term losses).48 While DFAX is not a perfect application of the beneficiary-pays principle. to capture extra rents. either in the even highly beneficial new investments if some short or in the long term. the cost beneficiaries is accompanied by less uncertainty of long interconnection lines is charged 100% about others. increased costs too widely may reduce cost discipline and reliability. all costs levied on tors. so tainty exists in the estimation of beneficiaries.

Transmission is inherently about not depend on individual commercial trans- moving electric power between locations. Instead. generators the benefits. for instance. and information on the identities of the beneficia. generator does not build Cost Allocation by Beneficiary Pays Generator Charge = 50 Load Charge = 50 Outcome: Generator charge < benefit. pre-arranged costs) over additional network costs. Because commercial Conceptually. This approach any subsidy to renewables and the decision of can yield a workable approximation to a how much to charge these generators as a beneficiary-pays allocation and make cost function of their economic benefits. allows for estimation of cost shares entirely bears the cost of other network rein. Figure 4. in which. generator builds and all network users are better off All this stands in contrast to some current generation and load. any network constraints. while load reliability. Transmission charges should be look for investments with the largest margin independent of commercial transactions. of aggregated benefits (or reduction in system Regardless of any specific. with careful consideration of the are at least initially responsible for the entire geographic scope of benefits from increased cost of radial interconnection lines. A sound commercial electricity trades.1 Illustrative Example for Generator Cost-Allocation Outcomes Generator Load Transmission Line Total Benefit = 120 Generator Benefit = 60 Load Benefit = 60 Line Cost = 100 Unprincipled Cost Allocation Generator Charge = 100 Load Charge = 0 Outcome: Generator charge > benefit.50 Any transmission planning exercise should Principle 2. Regulators should be aware of the respecting the principle that those in regions link that exists between the economic value of who do not benefit do not pay. transmission charges should the analysis of the value of such investment depend only on the location of the network requires calculation of locational impacts on 92 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . socialization a last resort. the physical flows planning process must provide sufficient on the network will remain unchanged. charges for network use should principle. loads will always be served by the least-cost ries of proposed transmission investments to set of available generators that does not violate enable those proposals to be evaluated. that make the beneficiaries better off while forcements. and actions. A consistent parsing of procedures. this information can be used to transactions have no influence on the physical allocate costs according to the beneficiary-pays network flows.

that argument does not rationalize the wholesale distortion of When planners fail to separate transmission transmission charges that pancaking implies. This that residents of C will bear the environmental second principle is not tantamount to socializa. Moreover. regardless of actual of anticipated benefits. but not in inter-ISO trans. While one can argue that compensa- location and the timing of network utilization. Furthermore. inefficient transmission investments and would significantly complicate operations in networks. a transmission investment ex ante. the result can be pancaking. charges from commercial transactions. in response. impact of the line and thus are entitled to tion of network costs. the net benefits as the difference in expected FERC issued Order No. Consider a line connecting regions A and B that crosses region C but According to this second principle. and consumers in A and B who benefit from serving entity in a region B should pay the same transmission charge as if. but access to the transmission system and the EU there is no comparable method for ex post developed a standardized mechanism for evaluation of benefits.users within the system and on when and where One possible argument for pancaking is that it power is injected and withdrawn from the can enable compensation to some losers from a system. ISOs. and C have are independent of commercial transactions been built in order. for instance. linking that project’s investors have already responded to benefits calculations and contracts creates the signal and committed to action. transmission charges generally power flows. The generators located in a region A that trades with a load. The existence of any contracts voluntarily signed by and on when and where power is injected and any agents should not affect application of this withdrawn from the system. There is borders between buyer and seller.S. The possi- perverse incentives for entering into contracts bility of future updates adds uncertainty and to avoid cost allocations. This could lead to raises capital costs. transmission charges left in place for the life of the project—or at least depend on the number of administrative a long period on the order of a decade. as indicated before. A feature of network accessing and paying for the transmission interactions is the strong interdependence of system. tion is justified in this case. Such pricing no reason to update a long-term price signal tends to stifle trade and prevent buyers from soon after a project has been completed because accessing low-cost sellers. no sense to consider benefits ex post in a actions. B. This principle should also be applied in hypothetical network with B and C but without inter-ISO transactions. that allocation should be power flows. a generator provides no benefits within it. Once costs have been allocated on the basis contract to pass through.S. A because as a general matter the presence of Chapter 4: Transmission Expansion 93 . principle because they should modify neither the physical real-time efficient dispatch of the line should pay for it. 888 providing open benefits with and without the investment. and the European Union have recog. As a result. it is possible and necessary to evaluate The U. After projects A. The allocation of costs should network users are required to pay accumulating be established ex ante. defining nized that pancaking is undesirable. it makes within U.52 Today. before the project is fees in every region their power is deemed by built. a situation in which Principle 3.51 transmission project. instead. One might argue generation nor the pattern of demand. it were Transmission charges should depend only on the contracted to supply a neighboring load sited location of the network users within the system within its own region—and vice versa. charge for the transmission of electricity transmission charges should depend on the through it.

the lumpiness and economies of scale period on the order of a decade. r $PTUTTIPVMECFBMMPDBUFEJOQSPQPSUJPO ex post calculation is neither easy nor required to benefits. when allocation methods in some interregional they are not operating at full capacity. that allocation should be left in uncertainty regarding future benefits. This may uncertainty is resolved rather than risk provide economic justification for at least imposing costs well in excess of ex post benefits partial socialization of many projects during on some parties.53 In place for the life of the project—or at least a long addition. as when significant uncertainty about future benefits is expected to be resolved is built. Current transmission cost. nor is there a principled framework for parsing benefits in r 5SBOTNJTTJPODIBSHFTTIPVMECF an interconnected grid. the early years of their operational lives. it may be sensible to socialize the costs for the early years of opera- Applying these principles perfectly is difficult tion until those users appear. and the resulting pattern of invest. r 5IFBMMPDBUJPOPGDPTUTTIPVMECF Applying this principle sometimes requires established ex ante. in practice. principles of cost allocation form a sound ments is likely to be inefficient. Ex ante To achieve an efficient and reliable network. and the same calculations can support ex ante cost allocation. such as in Central America and the portion of capacity is put in place to serve EU. A will have affected the designs of B and C FINDING as well as the decisions to build them. in the relatively near term. For example. if they are roughly compatible with voluntary the final system of cost allocation will lack agreements by the parties involved. the imminent location of a large generating plant might have a material impact on the expected Cost Allocation in Practice distribution of benefits. But unless one begins with sound principles and departs from them only to the Cost-allocation procedures are only workable extent required by practical considerations. In such cases. calculation of expected benefits is necessary three cost allocation principles should be for analyzing a decision to go forward with followed as closely as possible: an investment. By contrast. potential future users. If a markets. have adopted this approach. foundation on which mutually beneficial agreements can be constructed. Recognizing that the first of them—the beneficiaries-pay principle—can have somewhat different operational meanings in different settings. 94 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . of transmission lines means that it is often sensible to build facilities with excess capacity be most sensible to allocate cost after the in anticipation of future needs. The three coherence. it may In practice. before the project judgment. especially because of anticipated benefits. the process of determining who benefits and how much they benefit is Once costs have been allocated on the basis of analytically complex. independent of commercial transactions. for sunk investment decisions.

An interesting example of such a proce- recently built lines. s -AINTAIN CURRENT TRANSMISSION NETWORK The former might emerge from discussions charges for existing lines. and apply the among planning regions or be imposed by cost-allocation method only to new or FERC.we present some very general guidelines for interconnection-wide procedures or a set of allocation of intraregional transmission bilateral or multilateral cost-allocation agree- network costs: ments within each interconnection.55 This mechanism uses network s &OR ANY NEW POTENTIAL GENERATOR.x dure is the European Inter-TSO Compensation Mechanism.

The result could be either common x Our major concern is the allocation of the costs of new transmission investments. FERC might Allocation. operators (TSOs in European terminology). but some practical implementation guidelines can be found in the literature. and its network users pay based on that region’s chosen tariff method. loop flows. This hierarchical s 5SE THIS INFORMATION TO COMPUTE THE scheme provides a workable (if imperfect) transmission charges for recently installed interregional cost-allocation system anchored and prospective generators and loads. This outcome for intraregional projects. agreements among all affected parties. regional operator free to define its own system for intraregional cost allocation. Any existing transmission capacity expansion Planners first employ flow-based methodologies planning procedure should be helpful to determine how much external agents use the in the evaluation of the benefits of new network of each region. region. built soon have an impact on the network The net balance is credited or charged to each expansion plans. In this case. But they will become would obviously lead to processes that would more important in the future if large-scale be more sensitive to regional differences. alternative is a set of bilateral or multilateral tion projects are even less developed than those cost-allocation agreements. 1000 calls for the consider developing default procedures for development of standardized interregional multiregional issues in the absence of prior planning and cost-allocation procedures. Even gradual implementation of these guide- lines should move transmission tariffs toward On the other hand. such a system would be less adept at dealing with multiregional problems involving.54 ties involved in both the Eastern and Western Interconnections to reach agreement on this or Cost-allocation procedures for interregional. Transmission charges for the existing lines to new network users are of lesser importance to this study. for Interregional or Interconnection-Wide Cost instance. ESTIMATE flows as a proxy for benefits and deals only with the benefits of recently built lines and lines the allocation of costs between regional system that are scheduled to enter in service shortly. The interconnection-wide. and renewable genera. but renewable generation is substantially expanded. any other common interregional approach. it may be difficult as a a more cost-reflective—and thus more practical matter for the many regional authori- economically efficient—structure. and generators that the costs associated with that usage and allocate have been built recently or are planned to be them to the corresponding external regions. as well in beneficiary-pays logic and leaves each as for the remaining network users. FERC’s Order No.56 Chapter 4: Transmission Expansion 95 . Then they calculate transmission facilities.

First. Ideally. they should centers and remote renewable generators just as be applied globally within each interconnect: for other transmission projects. From these considerations emerge broad determine cost allocations for interconnections guidelines for sound interregional transmission of major system upgrades between major load cost-allocation procedures. s !PPLY COST ALLOCATION METHOD ONLY TO NEW Two other issues associated with renewables transmission projects. development deserve mention. wind and solar power plants are often built in relatively s 7HENEVER POSSIBLE.

That principle should be applied to anticipated to load. good resources that may ultimately support many such plants. which could take years. current tariffs who would pay for such lines. When all generators were built by vertically integrated A second problem emerges because many of utilities. USE AN ESTIMATE OF THE small increments of several tens to a few project’s benefits to allocate costs among the hundred megawatts. which should be built on In the meantime. different entities may now build necessary to accommodate these resources. however. could face the cost burden of oversized. did they have As a result. but only in unusual cases. to wait for generators to build first so that they can then finance the transmission upgrades In contrast. like mine-mouth coal plants. But in the case of remote problems. this is likely to represent a signifi. internal procedures. some than the standard sizes of conventional thermal measure of network usage might be used as a power plants. the generator would pay There are creative approaches to address both the full project cost. Initially. gigawatt of capacity or more. Under the traditional approach. since it is generally unclear under locational decisions. regions could allocate the renewables. utilities have little interest in building capacity Transmission cost differences were thus to remote areas far from the existing infra- automatically taken into account in making structure. Transmission small) part of the cost of the generator. generators will not build if their and the cost of interconnection lines and other plants have to sit idle for years before they can required transmission network upgrades may interconnect to the grid and start to sell their represent a significant fraction of the cost of the power. transmission utilities would like much of an effect on those decisions. may be developed in the future. Remote Renewable Generators. High-voltage transmission s 5SE THIS INFORMATION TO COMPUTE THE FRACTION lines. and own generation and transmission facilities. Of course. under- utilized transmission system upgrades. lines connecting the generator to the the best wind and solar resources are far from transmission grid were treated as (a generally the existing transmission system. a relatively small generator the three basic principles presented here. The effect of this mismatch is that large amounts of transmission s !LLOW EACH REGION TO ALLOCATE THE COST IT MUST capacity may not be used until more generation cover to its network users according to its own comes online in an area. Then. as generators come 96 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . large-scale wind and solar plants that problem with transmission. This is true even in areas with proxy for benefits. typically substantially less regions involved. if this is not possible.57 This is the classic chicken-and-egg remote. are often most efficiently of the cost of the project to be covered by constructed at scales designed to serve a each involved region. cost of new transmission projects in remote cant departure from the beneficiary-pays areas where wind or solar development is principle.

projects are most vulnerable developing in-state renewable resources. In this way. Shortcomings in planning or cost allocation even paying back consumers if appropriate. city and county authorities also California has instituted procedures of this may be involved. The rules in one state may a headroom account whereby future developers specify requirements different than in another reimburse the developer who has initially paid that affects the same investment—e.63 This basic conflict is not unique to permits from some set of states. issue. of transmission that may accrue to neighboring mission line. their pro rata share of the transmission costs. 1000. And as Ashley C. in its processes for During the siting process. If one hopes to build a line across CAPACITY multiple states or utility systems. hierarchically federal authorities to build the facility. they would assume not satisfied with the project for any reason. can compound intrinsic difficulties with siting. planners can A major test of this hypothesis would follow reduce the financial risk to transmission from implementation of the new planning and developers.g. and the loads subsidize logistical troubles will undoubtedly persist to the costs.58 Texas to challenge and litigation by parties who are not is using its CREZ planning. This is not strictly compatible with some extent. localities.. will ever become routine: local protests and tion does not appear. called location-constrained resource interconnection pricing. new genera. and the U. It is hard to imagine.” 62 This mismatch in scope compounds siting 4. would not force generation investors to inaction.60 incentive to be parochial in siting decisions. at efficient scale. State laws and regulations primarily govern the with well-defined steps. siting regimes have not kept up with this among numerous generators and establishing expansion in scope.S. and a lack of available transmission cost-allocation proposals in FERC Order No. if generators come on-line as such as NIMBY (“not in my backyard”) forecast. While recent decades have sort.61 ex ante cost allocation. projects are most vulnerable to challenge and litigation by parties who are Chapter 4: Transmission Expansion 97 .on-line in these areas. Yet another a specific route in Iowa and a set of alternatives alternative approach is the coordinated in Illinois. requiring for the transmission upgrade. it must acquire necessary siting systems. to address this satisfied with the project for any reason.64 the siting process. Conversely. “There is a powerful economic the New England states are discussing.4 SITING NEW TRANSMISSION difficulties. though in some states. that siting A risk is that the forecast is wrong. but it does involve a process that would be put in place ex ante. but costs would shift over time to sion planning and cost allocation should serve ensure that initial financing challenges do not to reduce disputes about cost allocation that prevent utilities from building a beneficial line surface in disguise during siting proceedings. it appears in other large. approval process for siting transmission. model that Rossi note.59 The New York ISO also has dealt with this issue in its interconnection procedures by seen a steady movement toward greater trans- initially using a “class-year” allocation process mission grid interconnection and regionaliza- to share the costs of transmission upgrades tion. Brown and Jim procurement. though.. or anchor tenant. progress on transmis- all parties. Eventually. During organized power systems. limited provisions are in place to recognize the benefits When a developer attempts to build a trans. which involves socialized regional cost recovery. the proper costs would be allocated to complaints.

and federal agencies with a conser. recreational land. December 2002. New Jersey and Pennsylvania agencies. Arizona regulators reviews adhering to the National Environmental unanimously rejected Southern California Policy Act. approval timelines for certain individuals or communities objecting to the sections of a right-of-way may exceed those aesthetic or perceived health or environmental for others because of the differing siting impacts of transmission infrastructure. Forest due to expanded interstate commerce is a Service. because state percentages of several western states. in May 2007. The increased cost for some because of opposition by the U. and the Advisory of September 2011. and higher on the determination of need. the Council regulators have approved the project. the Edison’s proposed Devers–Palo Verde 2 line. The utility proposed a route in by a 2006 memorandum of understanding on August 2008 after a detailed study and public transmission siting signed by five executive workshops.68 federal executive agencies. NIMBY authorities.66 Within the has not. an additional 12 major subagency organizations are also cited Siting permits and environmental reviews as playing a significant role in the process. The line was energized in June passes a complex system of many stakeholders 2006. the Endangered Species Act. each with its own Electric & Gas Group’s proposed Susquehanna– interests and rules. Obtaining regulators are often required by political approval to build a transmission line across reality—and sometimes by law—to focus federal land is never an ordinary commercial exclusively on in-state costs and benefits in transaction. which would link Pennsylvania involved in transmission siting is well illustrated and New Jersey. and other laws. Migratory Bird Treaty Act. undergo federal and state environmental For instance.projectnoproject. The current siting landscape in the U. 98 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .S. In March Arizona to California. Largely increased demand. the National Park Service Council on Historic Preservation.S. Chamber of Commerce maintains a useful resource that describes many of these challenges and includes a list of active transmission projects facing major hurdles (http://www. two regulatory bodies. securing an Nationally.67 Another example is Public Service and administrative processes. final approvals were not received until normal consequence of trade. Even without xi The U. and scenic or historic trails and parks have intensified. encom. often have a limited shelf life and thus may lapse if other complications cause delays. In At the local level.xi fragile ecosystems.69 The requirements to give vation mandate can strongly resist the priority to local need can be problematic even construction of high-voltage transmission when there are stakeholders in a state that facilities. but as on Environmental Quality.S. challenges take the form of other cases. which one commissioner described as “a 230-mile extension cord” pulling energy from Challenges can take years to resolve.70 They found that 1990. Groups opposing a project can opposition has grown over time as concerns for exploit the mismatch between processes.65 The plethora of authorities Roseland project. making decisions.com/ category/project/transmission/). the federal government controls authorization for a transmission project hinges about 30% of land in the U. Even on land that is not federally might benefit from the local development and controlled. American Electric Power announced its California ratepayers would benefit from access intention to build a 765 kilovolt line between to Arizona’s generating capacity. while Arizona Virginia and West Virginia that would pass rates would increase as a consequence of the through the Jefferson National Forest.S. Perhaps more than anything else. projects to build new lines must tax base of the new transmission investment.

of renewable power across long distances and across ple. For exam. If national and being voluntary present real problems policy is best implemented by bulk transfer of fashioning agreements that compromise of renewable power across long distances and different interests of competitors. the focus on local If national policy is best implemented by bulk transfer need is an obstacle to development. For example. many transmission though no such compacts have been formally projects have been planned. the Energy Policy Act significant impediment. Chapter 4: Transmission Expansion 99 .72 A bill siting provisions. approval for more than one year after the filing of an application seeking approval.71 In 2009 the House of Representatives Association established a protocol in 2002 to passed a bill with a provision that would have set forth procedures for collaboration between empowered FERC to consider interstate siting agencies in the Western Interconnection. and built in recognized yet. the Western Governors flawed. could preclude a required demonstration of meeting a need under state law. Despite these impediments. a transmission line that crossed a state like Arkansas without local interconnections might state lines. authority within NIETCs. legally designated NIETCs within which it could exercise that authority. similar organizations exist in the Midwest applied in both Eastern and Western Inter- ISO. and state authorities are responsible for which gave FERC authority to issue permits approving them. of 2005 empowered groups of states to form interstate compacts for transmission siting. but the absence of local service will likely prove to be a significant impediment. connections. it may serve as a basis for reported out of the Senate Committee on more detailed and binding future steps. ruling that lines that serve interests in multiple states. though While this agreement did not contain specific only in the Western Interconnection. there now exist no These regional efforts are still works in progress. These will tend to affect all states within a region. including than withholds approval of a project it opposes interstate cooperatives and joint transmission and that the process to designate NIETCs was studies. Energy and Natural Resources contained a Although differing in their details and execu. this act also recent decades. and PJM with the purpose of under. lines. broadly similar provision that would have tion.affecting electricity prices. standing and coordinating transmission siting and even if FERC had meaningful backstop processes across the different states involved. the existing state-centered siting procedures make sense. some states have gradu. simply because reliability problems Electric Transmission Corridor (NIETC). projects rejected by state regulators. across state lines. the existing state-centered siting procedures will likely prove to be a At the federal level. subsequent Circuit Court decisions ally taken action to ease the siting process for made §216 effectively irrelevant. However. They FERC cannot act if a state simply rejects rather have employed several mechanisms. permits would confer rights of eminent domain Problems arise when projects serve economic if a state commission or other entity with or public policy goals and involve costs and authority to approve siting has “withheld benefits in multiple states or regions. More notably. Many of these do not cross state added the new §216 to the Federal Power Act.73 Neither provision became law. the Secretary of Energy as a National Interest ations. ISOs’ planning processes have for facilities in areas experiencing capacity also been successful in facilitating multistate constraints or congestion and designated by lines that are justified by reliability consider.” As regional institutions and processes have grown in importance. sited. SPP.

to obtain permission to build a transmission project located in a single state that does s %STABLISH OR USE EXISTING REGIONAL INSTITUTIONS to facilitate interstate siting and standardize not cross federal lands than one that siting procedures. Local authorities held sway over need a centralized siting agency. increasing interconnec- transmission facilities on federal lands.75 tion and even broader regional and national goals—but now the role of individual states. not to mention local governments. state governments began a slow and incomplete standing signed in 2006 is an example of process of centralizing decisions within the action in this direction.76 The early technology for electric transmission s $EVELOP COMMON REVIEW PROCESSES BETWEEN emphasized local development of generation local. State and federal agency officials may s %XAMINE STATE LEGAL FRAMEWORKS TO IDENTIFY be doing their assigned jobs well yet acting legislative language that may inhibit siting against the broader national interest. including and limited (or no) longer-distance intercon- coordination of requirements and potentially nection. they are relatively to streamline the siting process and lower modest and unlikely substantially to reduce unnecessarily high barriers to transmission obstacles that arise because responsible state development: 74 agencies serve the interests of their in-state constituents at the expense of others and s !RTICULATE BEST PRACTICES FOR REVIEWING SITING federal agencies pursue mandates that give little proposals. all in announced formation of an interagency pursuit of the broader state-wide interest. As conditions changed and the impact of trans- s #OORDINATE AND SPEED FEDERAL AGENCY REVIEWS mission grew to cover larger areas and loads. A range of other measures has been proposed at least in part. An analogous problem arose within states in public policy. This coordination and how that can be remedied. The Renewable Energy Rapid Response Team same technical process continues with longer- to ensure timely review of proposals to site distance transmission. Because existing transmission siting procedures are widely recognized as a hurdle to develop- ment. as is the recently state and preempting local authorities. The interagency memorandum of under. state. The encompass the broader interests of larger regions or parochial interests of the states and localities. determination and the siting process. has changed which are easy to understand. 100 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . The parochial interests of the states and localities. requires the approval of more than one state or a federal agency. s "ROADEN THE DElNITION OF NEED THAT STATE commissions use to include energy efficiency. do not naturally against the backdrop of this broader reach of proposed transmission investments. FINDING s )NCREASE REGULATION OF THE TIME ALLOWED FOR Current siting procedures make it easier permitting reviews at all levels. and out-of-state benefits. of the nation as a whole. which are easy to understand. the early years of the grid’s development. weight to an efficient and reliable bulk power system. and federal authorities. some of these reforms may be realized. However. do not naturally encompass the broader interests of larger regions or of the nation as a whole. structural problem is unlikely to be solved without effective structural change.

S. Implementing a transmission expansion has served the U. One might want to depart in Order No. and interconnection-wide analysis The current system for governing and financing and planning would help. it will likely become more RECOMMENDATIONS important in the future. anywhere in the pipeline. Improving regional. while specifying that a state’s denial of administered. beneficiary-pays cost-allocation electric power industry well. Because some of the best wind and interstate commerce. interregional. The simplest and most elegant solution to the problem of interstate transmission siting is to These processes and associated institutions give FERC authority over significant interstate could be required by federal legislation. permanent. It is important now and. allowing FERC to consider negotiated terms. as well as economic and public policy not be optimally designed or perfectly benefits. when large deployment of the competing interests of states would affect grid-scale wind and solar generation is antici- electricity transmission in the same way that pated and the transmission grid becomes more conflicting state interest could have affected interconnected across state and regional natural gas pipelines and other forms of boundaries. interregional.We have also seen this process before in the authority to projects that have emerged from case of natural gas.S. as appropriate. or projects or those requiring land managed by FERC could extend the planning requirements another federal agency.77 While this procedure may reliability. In 1938. their efficient utilization will require desirable investment is not enough. this procedure within each interconnection would system likely will not be adequate in the help. Congress regional. Since a 1947 amendment to this Act to give the FERC effective backstop Section. hoping that solar resources are located far from major load obvious structural problems will not retard centers. the company building it receives the U. While one can debate the merits of these alternative approaches—and we have done so The problem of siting interstate electric within our study team—we agree that either transmission facilities was not important in the natural gas model or the FERC backstop 1935 when the Federal Power Act was passed. However. approach would serve the national interest and it was not addressed in that legislation. immediate future. effective interconnection-wide transmission planning processes. This would involve eliminating the right to acquire the necessary property by conceptually and administratively troublesome eminent domain if it cannot acquire it on notion of NIETCs. much better than the status quo. In the end. as we have argued. if the FERC approves a proposed authority over these projects. it does not generate either approval of a multistate project should serve as intense controversies or unreasonable delays a trigger to FERC consideration. compared to the analogous challenges of electricity transmission expansion.5 CONCLUSIONS AND of remote renewables. or interconnection-wide recognized that a similar structural problem planning processes. Section 7(h) grants of local conditions. workable. in large part because of the growing importance 4. Yet elimi- could substantially retard the development nating states’ roles entirely has a variety of of interstate natural gas pipelines and passed disadvantages given their superior knowledge the Natural Gas Act. 1000 from regional and bilateral from the natural gas model by limiting FERC’s interregional levels to the interconnection-wide Chapter 4: Transmission Expansion 101 . But even with perfection in these reforms. Thus a workable alternative FERC authority to evaluate proposals for such would be to amend §216 of the Federal Power pipelines.

available data and R E CO M M E N D AT I O N planning methods do not yet support rigorous interconnection-wide planning that adequately The industry. comprehensive approach U. and in recent years the Census Bureau and other a consistent approach to all is required to federal agencies have developed protocols for produce stakeholder support for an efficient making highly confidential data available to and reliable system. procedures that satisfy security concerns. and siting are all interrelated. improve utilization—of better planning methods is system reliability. an ad hoc industry coalition. the industry could see the logic detailed and comprehensive data on the and value of a broad. level. business models. the criteria and We believe similar protocols can be devised for decision-making process that are adopted bulk power system data. or a public-private partnership involving DOE. NERC’s mandate To produce coherent outcomes. particularly in the both should support research to improve more complex Eastern Interconnection. In particular. It could be funded by industry transmission projects at the interconnection contributions to the Electric Power Research level. Order No. bulk power system and make them to planning and voluntarily go beyond the available to researchers and others under order’s minimum requirements.S. cost would obviously raise security concerns. Planning. A responsible entity should develop Alternatively. Unfortunately. Institute. we do not believe the necessary processes should be established for research is likely to be expensive by the stan- conducting planning of interregional dards of federal energy research and develop- ment projects. regulation has to be a conceptually integrated Making detailed network data widely available system. robust methods for wide-area transmission planning under uncertainty The first need is to develop the necessary data over multiple time periods. for transmission planning need to be closely coupled to the subsequent cost-allocation process. This does not require the creation of a new agency. 1000 recognizes. as FERC researchers without compromising security. the federal government. but allocation. 102 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . hierarchical. and increase efficiency. or accounts for uncertainty. and make them available to those who can use them constructively. R E CO M M E N D AT I O N To support the integration of large-scale While we believe the development—and wind and solar resources. leaving the details of procedures and R E CO M M E N D AT I O N organization to internal industry agreements. important and an attractive area for academic permanent hierarchical and collaborative research. however. transmission could be broadened to include this function.

but the challenge can be met through and establish one voluntarily. No. it is These incentives constitute barriers to siting important to develop an agreed-upon proce. agreement with the three principles. they should be compatible with planning and cost allocation will make it easier. integration of large-scale renewables. Even with these improvements.R E CO M M E N D AT I O N R E CO M M E N D AT I O N In the interest of producing reliable A hierarchical approach to interregional electric power at least cost. and used to calculate expected benefits to the task of cost-allocation methods must enable trans- displaying the expected distribution of benefits mission projects that have justifications other that are inherent in the analysis of transmission than reliability to be built under appropriate projects. FERC’s Order No. interconnection would have a single procedure rather than a maze of bilateral or multilateral agreements. planning ance filings should apply the same information criteria. While intra. 1000 sets out the advance overall system efficiency and efficient core principles. and the allocation of costs should in business models. Siting transmission facilities will always be regional cost-allocation methods may differ a complex issue. rules for interregional network cost out-of-state interests and for managers of allocation do not exist. health and efficiency of the bulk power system. in particular by cross regional boundaries or have a significant providing for the efficient integration of impact on interregional trade. connectedness of the transmission system. Those responding with compli. also in as practical in proportion to benefits. but sound approaches to somewhat. transmission cost-allocation respecting the three principles of the methods should abide by three basic preceding recommendation. decision-making procedures. this goes r 5IFBMMPDBUJPOPGDPTUTTIPVMECF beyond the requirements of FERC’s Order established ex ante. application of the same tools used to perform cost–benefit analysis in transmission expansion If investments in the transmission grid are to planning. to allocate principles: costs between regions. Like the first recommendation. however. these three principles. each large-scale wind and solar generation. principle be exactly proportional to the expected distribution of benefits. Federal legislation or a new FERC order could mandate this approach. interstate transmission facilities that serve the dure for allocating the costs of projects that broader national interest. generation. Each region then can cover the costs allocated to it using its own r $PTUTTIPVMECFBMMPDBUFEBTDMPTFMZ internal cost-allocation procedure. r 5SBOTNJTTJPODIBSHFTTIPVMECF independent of commercial transactions. Or the industry could see the advantages of having a Implementation of these principles will be a single interregional cost allocation procedure challenge. there are strong incentives for state agencies to ignore Currently. Thus. particularly transmission cost allocation should be while integrating large-scale renewable developed that uses a single procedure. 1000. Chapter 4: Transmission Expansion 103 . Ideally. to accommodate federal lands to give inadequate weight to the renewable generation and the growing inter.

S. Such authority could broadly parallel its authority over interstate natural gas pipelines. these two approaches have different strengths and weaknesses. but either would be a significant improvement over the status quo. As we have discussed above. 104 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . or §216 of the Federal Power Act should be amended to give the FERC effective backstop siting authority anywhere in the U. R E CO M M E N D AT I O N The federal government should grant FERC enhanced siting authority over interstate electricity transmission projects or those that cross land managed by another federal agency.

Economics. Pfeifenberger. 1000 (July 21. “Potential Coal Plant Retirements Under Emerging P. 2009). DC. midwestiso. Calgary. D. 2009. Tennessee Valley Authority. “21st Century Perspective: On Allocating the Costs of New Transmission Planning: The Intersection of Transmission Investment: Practice and Principles.” Public Wholesale Power Markets. see note 10 above. 20% Wind Energy by Kwok. Wilson. Transactions on Power Systems 10. http://www. Kwok. MA: MIT Center for Energy and Environmental Policy Research.hks. 2009). 2011).misoenergy. CIGRE/ Systems (Washington. 4 Federal Energy Regulatory Commission. July 11 Edison Electric Institute. see note 4 J. and T. 2011). 7 Coalition for Fair Transmission Policy. DC: National Governors Association. and R.” working paper 2005-004 (Cambridge. Celebi. 2 M.” The Brattle Group. Marín. 1 (1995): 2009).” Engineering. IN. DC: U. AR: org/Library/Repository/Study/MTEP/MTEP09/ Southwest Power Pool Communications MTEP09%20Report. Hirst. Canada. Paper 978-1. The Benefits of 2009 (Carmel.dsireusa. Kaplan.” IEEE (Washington. Minneapolis. and Environment. J. 2010). 546–553. Mid-Continent Area Power Pool. Electric Power Industry (Washington. https://www. and Joskow and Tirole. Midwest ISO Regional Generation 17 3 Energy Information Administration. see note 14 above. Rubio. Puerta.brattle. 19 2030 (Washington. PJM Interconnection. http://www. Investment. Joint Federal Energy Regulatory Commission. Bathla. http:// J. 26 29–31. Mount. R. above. (Cambridge. Washington. Department of Energy. F. 2011). “Barriers to Transmission 24 above. 16 asp?PublicationID=1243. Congressional Research Service. April 12. http://www.” 14 Public Utilities Fortnightly.” Proceedings of the IEEE Investment. Wilcox. Publications/ReportsPresentations.pdf. 2010 General Meeting. DC. http://www. DC: Power Pool. FERC-11-R-1011 (Washington.S.nrel. Order No. H.org/. S. J. see note 10 above. 14 (2000): 54. Outlook 2011 (Washington.harvard. and Southwest Power Pool. Pérez-Arriaga.” presentation to the Utilities Fortnightly 138. and Kwok. Springer and G. J. Kwok. J. “Patterns of Transmission Old World Meets the New. L. G. 2009). “Transmission Crisis Looming. 2011). DC.thecftp. Statistical Yearbook of the 2010.ferc.pdf. “Marginal Pricing of Transmission An Infrastructure Vision for the 21st Century. MA: Harvard University. and L.” white paper Transmission Owning and Operating Public Utilities. I. and FERC. April 2010. “Electricity Transmission Investment in of Non-Traditional Transmission Development” the United States: An Investigation of Adequacy” presentation to IEEE Power and Energy Society (Master’s thesis. “A National 25 12 R.spp. DC. 23 www.com/ (2005): 233-264. 1 Database of States Incentives for Renewables Economic Assessment of Resource Adequacy & Efficiency. 2011. M. “Reconsidering Resource Adequacy. Blue Ribbon Panel on Cost Allocation. July R.pdf. 2010). L. MIT. http://www. Joskow. a Transmission Superhighway (Little Rock. MN. 18 5 U. 2008). Department. aspx?id=18974. “Transmission Transmission Planning and Cost Allocation by Benefits and Cost Allocation. Services: An Analysis of Cost Recovery. 11 (2005): 2026–2035. no. Electric Power Transmission: 21 6 Midwest ISO. 93. Dierkers. F. Annual Energy Outlet Study (Carmel.REFERENCES J. https://www. and W. IN. and J.edu/hepg/Papers/2011/ gov/industries/electric/indus-act/trans-plan/ Hogan_Trans_Cost_053111. fr-notice. American Antitrust Institute. J. J. DC: WIRES.” white paper prepared for the Working group for presented at Integration of Wide-Scale Renewable Investment in Reliable and Economic electric Resources into the Power Delivery System. and D. Coxe and L. MISO Transmission Expansion Plan 20 pdf.” Journal of Industrial Economics 53 December 8. no. Whitehead.org. Arceluz. Gutman and E. Midwest ISO.F. http://www. 2011). Chapter 4: Transmission Expansion 105 . Bressan. see note 4 22 Coordinated System Plan (2008). Southwest Background and Policy Issues (Washington. Meeus. 11–13.org/_layouts/miso/ecm/redirect. Midwest ISO. 2007). Graves. Joskow. F.gov/docs/fy08osti/41869. no. Thomas. “Transmission System Planning—The 13 P. 2010. Tirole. Joskow and J. Investments and Implications for Competition in 9 E.org/. “Merchant Transmission 15 Environmental Regulations. Hogan. Outhred. 8 Federal Energy Regulatory Commission. IEEE PES Joint Symposium.S. “Survey 10 P. see note 10 above. Requirements. 2005). T. see note 13 above. Department of Energy.

and PJM Inter- connection. “Guidelines on 51 Corporation. Glachant. Rubio and J.PDF. D. Department note 33 above. 2010). PA. M. edited by François Lévêque et al. 2010. 2005). 2009). A Study on the 49 Program on Energy Policy. and Southwest Power Pool. Hogan. http://www. Hogan. http:// www. Belgium.energy. U. J. http:// Allocation Methods. Final Report (Little Rock.” FERC “Transmission Network Investment as an Docket No.S. 35 V.” FERC Docket No. PJM Regional Transmission Expansion Planning Process (Valley Forge. NJ. R. Areiza. C. please see note 21 above. PA. Renewable Energy Transmission Planning: Process Document Transmission Initiative: Phase 1B. Security. Schweppe. California. Renewable Energy Transmission Initiative 41 2008). “Response of PJM Interconnection.” IEEE Transactions on Power www. 2008). EL05-121-006. Smeers. Review of Existing Kaplan. January 2009). May 28. 28 Western Electricity Coordinating Council. Highways System 2050: Way to 2050 Pan-European PJM Interconnection. 2010). see note 27 above. Dessante. Eastern Wind Integration ENTSO-E. ITCStudy-FinalReportCover_000. Department of Homeland BOD/TEPPC/default. Cruz. Hogan 2011. http://realisegrid. Systems 15. Netherlands: Kluwer Academic Publishers.com/filez/ivgtf.S. North 39 and Transmission Study (Golden. “Classification of Publications and Models W. https:// Cost Allocation Issues. above. A Survey of Transmission 42 Power System (Brussels. 106 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . 45 European Grid Towards 2020 Challenges and E.Villegas. “Marginal Pricing of Transmission Corporation.biz/committees/ (Argonne. 44 33 North American Electricity Reliability Pérez-Arriaga. and Public 2010–2020 (Brussels.eui. Inter-TSO Compensation Mechanism (Florence.entsoe.C. 2 (2003): no. M. no. I. (Sacramento. “Contract Networks for Electric Power 46 on Transmission Expansion Planning. and A. 938–946.it/. 3 (1992): 211–242.” Journal of Regulatory Economics 4. D. 2010/04 (Florence. Study Roadmap towards Modular Program Information Center. Pérez-Arriaga and Y. see note 17 above. see Renewable Energy Laboratory. see note 4 43 32 Federal Energy Regulatory Commission. R. 2010). the Netherlands: Kluwer Academic Publishers. Tabors. Italy: Loyola De Palacio Florence School of Regulation. Department of Energy. CREZ Transmission 31 ENTSO-E.nerc.wecc. 47 Methods for Transmission Planning and for Grid Connection of Wind Power Plants. (Dordrecht. Spot Pricing of Electricity (Dordrecht.Bohn. National 40 “Transmission Expansion Planning Policy Power Grid Simulation Capability: Needs and Issues Committee. see note 5 above. Research and Development Plan: F. Rious.eu/RSCAS/Research/FSR/pdf/051031- EnerNex Corporation. CO: National American Electric Reliability Corporation. 2009). Tariff Setting. Anticipation Problem. 37 Midwest ISO. U. 1 (2000): 448–454. Belgium.” http://www. see note 16 above. Department of Homeland Security. Corporation on Remand. I. 34 ENTSO-E. EL05-121-006.S. Science and Technology Directorate. Belgium. Ten-Year Network Development Plan 2008-003/RETI-1000-2008-003-F. and REALISEGRID. electricity-highways/.eu/system-development/2050. J.gov/2008publications/RETI-1000- 30 ENTSO-E. Accommodating High Levels of Services: A Comparative Analysis of Network Cost Variable Generation (Princeton. J.” in Transport Pricing in Electricity Markets. Transactions on Power Systems 18. see note 34 of Energy. Caramanis.ca.aspx. 2010). see note 31 and 34 above. Federal Energy Regulatory Commission. Latorre. and R. Utility Commission of Texas.” IEEE Transmission. and REALISEGRID. Italy: European University Institute. 2010). 29 PJM Interconnection.html. and P. see note 17 above.S. see note 33 above. 2010). 50 38 North American Electricity Reliability J. and F.” working paper EUI RSCAS. AR. Integrated Coordinating Committee. IL: U. Methods and Practices www. chapter 7. to Information Requests by FERC.com. EU 7th Exelon.rse-web. 1988). (Valley Forge. April 13. and www. 27 EnerNex Corporation. the Beyond (Brussels. above. see note 4 above. Development Plan on Pan-European Electricity texascrezprojects.L. “Initial Comments of EXELON 48 Framework Program Research Project (2009).pdf. 2003). http:// 36 U. 2011). G. L.

Belgium.” Siting in the Western United States: Overview and Office Journal of the European Union 250 Recommendations Prepared as Information to the (September 24. “Consultation Document 55 on the Inter-TSO Compensation Mechanism and Public Service Electric & Gas Group. A. and Holdkamp and Davidson 2009. Fox-Penner. aspx. DC. AZ Corporation Commission. Strategic Options for Investment in 57 May 30.S. Chapter 4: Transmission Expansion 107 .” white paper (Salt Lake City. 3 (2010): 705–711. DC. see note 65 above.com/ DG TREN/C2 (Brussels. Olmos and Pérez-Arriaga. no. 52 National Council on Electricity Policy. Pfeifenberger and P.” press release. Sedano. 5–11. N. 2009.” http:// 67 no. cert.. 2009). 1138 California ISO. S. American Electric Power. Rossi. ” FERC Docket U. A. AD09-8-000. Reider. and R. Recovery of Stranded Costs (Washington. www.pseg. DOE. Wasserstrom and S. Johannes Pfeifenberger. Holdkamp and M. 631 F.” in National Transmission Grid R. I. Relating to the Inter-transmission System Operator Compensation Mechanism and a Common J. Electricity Transmission Infrastructure and Its “Commission Regulation No 838/2010 of 23 Impact on the Renewable Energy Market September 2010 on Laying Down Guidelines (Washington. 2010). “A 54 in Order to Site Electric Transmission Facilities” Comprehensive Approach for Computation and (Washington. MA: Analysis Group. Federal Energy Regulatory Commission. no. Fox-Penner. 2006). and European Commission. 2008). Krapels.” http://www. Davidson.azcc. 72 Public Utility Commission of Texas. 2010). Eggs. Implementation of Efficient Electricity Transmission Network Charges.S. 63 Promoting Wholesale Competition Through Open Coordinating Interstate Electric Transmission Access Non-discriminatory Transmission Services Siting: An Introduction to the Debate by Public Utilities. “MOU on 66 No. Meyer. and Carbon Reduction (Austin. American Clean Energy and Security Act of 2009. Western Interstate Energy Board. Belgium. 1 (2010): 34–38).” “Susquehanna–Roseland. 6 (2009): 86–87. 2007. no. A Blueprint for E. Senate Committee on Energy and Natural 73 J.Ct. CO: Holland & P. Allocation. National Council on Energy Economics. 558 F.0 (Folsom. UT. 3d 1072 (9th Cir.R.” The Electricity Journal 22.” The Electricity Journal 23. and Interconnection Animus. December 18. European Commission. 71 3d 304 (4th Cir. A. World Resources Institute. DC. C. and family/pseandg/powerline/index. see note 63 above. “Summary: American Clean Energy Tenant Model—And Some of the Chickens and Leadership Act of 2009” (Washington. Version 7. Transmission Planning Process.pdf. 2009). “Transmission Siting 74 and Permitting. and Denver. Pérez-Arriaga. TX: Center for Department of Energy. 111th Congress (2009). 2010). see note 41 59 H. Olmos and J.aep. (2009).gov/divisions/ Transmission in Support of Offshore Wind administration/news/Devers_II_Vote. “The Terrible Trio Impeding 69 an Integrated European Energy Network. 68 on Harmonization of Transmission Tariffication. Business Practice Manual for the 58 (2010). Early Coordination of Federal Authorizations and Related Environmental Reviews Required L. and Delphine Hou. High Wire Act: 64 Order 888 (1996). DC. 2008). above. 12 (2009): 5285–5295. “Energy Infrastructure Priorities for 2020 and Beyond. D. “Transmission 65 Regulatory Approach to Transmission Charging. 53 Hart LLP. 2009). “Regulators Reject 70 ‘Extension Cord for California’.” Transmission Development: Siting. Development in Massachusetts (Boston. Electric Transmission 61 Study Issue Papers (Washington. Piedmont Environmental Council v. “The Anchor- 60 Resources. J. 2002). European Commission. Tierney. CA: California ISO.S. 2010). Hou.com/about/transmission/wj_chronology. and D.jsp. DC: U. and California Wilderness Coalition v. Pfeifenberger. by Public Utilities and Transmitting Utilities.” University of Colorado Law Review 81. 2009). “Chronology.” Energy Policy 37. Department of Energy et al. denied 130 S. FERC. 2011). 2454. Cost COM(2010) 677/4 (Brussels. 2011). http://www. U. “Siting Transmission 62 Lines in a Changed Milieu: Evolving Notions of the ‘Public Interest’ in Balancing State and Regional Considerations. “Comments of Peter Fox-Penner. H. see note 54 56 above. Electricity Policy. Brown and J.

whitehouse.pdf (accessed 9/8/2011). “Pre- Filing—FAQs. see note 66 above..gov/industries/gas/ indus-act/pre-filing/faqs. “FACT SHEET: The President’s Plan for a 21st Century Electric Grid. 2011). 75 U. 76 Brown and Rossi. June 13. and The White House. Department of Energy et al.” (Washington.” http://www.asp. see note 62 above. 108 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .ferc. 77 Federal Energy Regulatory Commission.C. D..S.gov/sites/default/files/microsites/ostp/ smart-grid-fact-sheet-6-13-2011. http://www.

These modifications are required to allow the full realization of several of the projected benefits of DG. these technologies are expected to increase in penetration over the next few decades. and bution primary voltage (13. Section 5. connected to our local utility at distri- (PV). fuel cells. It begins by introducing the different types of EVs and recent forecasts of their potential penetrations over the next several decades. Though not connected to the grid. Supported by public policies reflecting a range of concerns and goals. We find that influencing the timing of vehicle charging could improve system operation and avoid investments in infrastructure upgrades that would otherwise be necessary. including gas based on a gas turbine engine rated at about turbines. solar photovoltaics 20 MW.i For example. Similar to Chapters 2 and 3. At high penetrations they may require systemic changes in the way the electric grid is planned and operated. diesel engines. we do not discuss them. We introduce the primary interconnection standards for DG and discuss several potentially important modifications to the standards. 5. and power plant of generation technologies. wind turbines.1 focuses on DG. Some DG units that use conventional fuel-burning engines are Distributed generation refers to relatively designed to operate as combined heat and power small-scale generators that produce several (CHP) systems that are capable of providing kilowatts (kW) to tens of megawatts (MW) of heat for buildings or industrial processes using power and are generally connected to the grid the “waste” energy from electricity generation. Chapter 5: The Impact of Distributed Generation and Electric Vehicles 109 . We recommend that the main standard governing DG interconnection be revised to permit voltage regulation by DG units and that utilities provide incentives for off-peak vehicle charging in regions with high EV penetrations. we discuss the challenges and opportunities associated with distributed generation (DG) and electric vehicles (EV). Dispersed generation is typified by standby diesel generators that provide backup power in the event of a grid failure. dispersed generators can participate in demand response programs (see Chapter 7). The successful integration of growing penetrations of DG units and EVs primarily will be the concern of industry engineers.1 DISTRIBUTED GENERATION small hydroelectric generators. MIT.Chapter 5: The Impact of Distributed Generation and Electric Vehicles In this chapter. Distributed i It is important to note that distributed generation is distinct from dispersed generation.8 kV). It then describes the potential benefits of DG. It then describes electric vehicle charging requirements and discusses the importance of influencing the timing of electric vehicle charging. which is not connected to the grid.2 discusses EVs. we briefly describe several effects of DG on distribution system operations. this chapter provides important background and context for the chapters that follow. Section 5. cooling. followed by a discussion of the interconnection challenges related to DG. Finally. our own institution.1 at the distribution or substation levels. It starts by defining distributed generation and describes recent deployment trends. has a Distributed generation units use a wide range combined heating. Because these units typically do not impact utility operation or planning activities. Section 5. biomass.3 provides our conclusions and recommendations.

Ramakumar. However. The Potential Benefits of Distributed Generation and Rate-Related Issues that May Impede Their Expansion: A Study Pursuant to Section 1817 of the Energy Policy Act of 2005 (Washington. Department of Energy. wind. some states have Distributed generation can be owned and operated provisions in their renewable portfolio standards by utilities or their customers and can provide that require some fraction of retail electricity a variety of theoretical benefits to their owners and sales to come from renewable DG by 2020.S. Distributed generation In 2009. the largest installations or control the operation of small DG units. trial DG units with a combined capacity of reduce costs. Renewable DG from wind and solar power also Federal and state policies are expected to drive typically is not dispatchable or easily control. growth in DG in the coming decades.5 While 90% generation facilities do. and P. transmission. generated more than 14 MW. Chiradeja and R. the to utility systems in the U. generation can be owned and operated by engines. and and the broader power system. about 13. and steam utilities or their customers and can provide a turbines comprised more than 4 GW each of variety of theoretical benefits to their owners installed capacity. DC.3 Internal combustion characteristics of the local power system. Sixteen lable. no. have renewable portfolio standards with specific DG provisions. 93.000 commercial and indus.2 Of these units. installations theoretically can improve reliability.8 However.S. r 3FEVDFEIBSNPOJD congestion or distribution upgrades distortion r 3FEVDFEJNQBDUTGSPN r -PXFSPQFSBUJOHDPTUT physical or cyberattacks due to peak shaving r *ODSFBTFEHFOFSBUJPO r 3FEVDFEGVFMDPTUTEVFUP diversity increased overall efficiency r 3FEVDFEMBOEVTFGPS generation Source: U.000 residential PV installations with system operators like central station totaled about 450 MW of capacity. reduce emissions. Large DG units other generator technologies totaled 3 GW. 2007).7 For example.1). and improve about 16 gigawatts (GW) were connected power quality (see Table 5. neither of solar PV installations between 1998 and 2007 utilities nor system operators typically monitor were smaller than 10 kW. combustion turbines. the broader power system. 110 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . benefits of DG are highly dependent on the 10.6 especially those in residential applications. Table 5.1 Theoretical Benefits of Distributed Generation Reliability and Security Economic Benefits Emission Benefits Power Quality Benefits Benefits r *ODSFBTFETFDVSJUZGPS r 3FEVDFEDPTUTBTTPDJBUFE r 3FEVDFEMJOFMPTTFT r 7PMUBHFQSPñMF critical loads with power losses r 3FEVDFEQPMMVUBOU improvement r 3FMJFWFEUSBOTNJTTJPO r %FGFSSFEJOWFTUNFOUTGPS emissions r 3FEVDFEóJDLFS and distribution generation.” IEEE Transactions on Energy Conversion 19.4 In are typically dispatchable and communicate the same year. Distributed generation advocates cite a litany of good things DG can do. 4 (2004): 764–773. while hydroelectric. characteristics of each installation and the averaging 100 kW each. These units present the greatest challenge states and the District of Columbia currently and are the primary focus of this chapter.800 (83%) were smaller than 1 MW. “An Approach to Quantify the Technical Benefits of Distributed Generation.

uninterrupted power can improve action is called “islanding.11 At present installed costs. Combined heat and these costs are not competitive with conven- power (CHP) systems also can reduce total tional generating sources in most locations. and most wind turbines) are variable energy resources.12 The magnitudes of emissions benefits associated with DG depend on both the Improved system reliability results from the characteristics of individual DG units and the ability of DG units to maintain supply to local characteristics of the power system to which loads in the event of a broader system outage. or CHP the system. in some government policies will largely determine the cases. generators. it can reduce conges- tion and system losses in some instances. commercial.60 per Wdc generated in central stations. they are connected.15 As of September customers who use more than some amount 2011. Such an constant. and industrial of electric energy pay a high rate) or who are PV installed system costs had fallen to $7. many renewable Customer-sited DG. could theoretically cancel grid distortions and help regulate voltage. the power electronics operation is possible. On the other local loads and also the necessary distributed hand.10. existing DG interconnection systems whose use of waste heat can result in standards prevent owners from realizing some higher efficiencies than central generation of these hypothetical benefits.70 per Wdc .Furthermore. distributed generation connected to the system control capabilities. solar reliability benefits of generators based on PV. current rules allow several years. on the other hand. if designed and individual reliability are limited even if islanded implemented properly. units. installed cost of residential and commercial because electricity generated by DG installa. First.g.13 Many inverters on the Economic benefits can be realized when market today are capable of these advanced utilities deploy DG to defer investments in functions. respectively. such as solar photovoltaics (PV). if they continue to fall. However. and $3. generators with widely understood to be sources of voltage limited fuel reserves.50 per Wdc in 1998 to about $7. tion system operator or the other customers of which have no marginal emissions.10. Finally.9 The potential grid via power electronic inverters (e. This could be done by creating “islands” in which a section of a distribution feeder is Distributed generation capable of providing disconnected from a faulted area. and today transmission or distribution infrasturcture. residential.10 DG owners rarely have incentives to invest in Since DG is typically located closer to load this added functionality. solar PV systems will Chapter 5: The Impact of Distributed Generation and Electric Vehicles 111 . but such features add cost. The durability of such long-term electricity cost stability and. many benefits accrue to specific Emission benefits can be realized by renewable stakeholders and may not benefit the distribu. The average of fixed network costs (See Chapter 8).” Successful islanded power quality by mitigating flicker and other operation requires sufficient generation to serve voltage regulation problems. This savings can come in rate of growth of installations over the next different forms. Second. offered sufficient subsidies can realize energy $5. In the long term. energy costs for their owners.14 relative to central plants. savings. often DG installations remain dependent on these reduces utility revenue but can offer customers mandates or subsidies. solar PV installations dropped from about tions is typically more expensive than electricity $10.16 While cost savings with DG.. or generators with low waveform distortion. fuel cells. cost reductions customers with DG to avoid paying their share also may drive DG growth. customers subject in 2007 (both figures are in 2007 USD before to increasing block electricity tariffs (in which incentives or tax credits).

ultimately become competitive. As described other customers or equipment connected to
in Chapter 8, net metering policies that favor the grid; it applies to the interconnection
renewable DG could accelerate the adoption of all generation with aggregate capacity of
of residential rooftop solar PV generation even 10 megavolt amperes (10 MVA, approximately
before this type of generation becomes other- 10 MW) or less to the distribution system.
wise economically viable.
The standard includes several provisions to
FINDING mitigate DG’s potential negative impacts on
Distributed renewable generation, though power quality. For example, the standard
requires that DG not “create objectionable
becoming more cost competitive with
flicker for other customers.” 20 “Flicker” refers
conventional generation technologies,
to rapid variations of voltage that can cause
is still significantly more expensive and noticeable variations in lighting and interrupt
strongly dependent on mandates and the operations of electronics. Flicker can occur,
subsidies for its economic viability. for example, when clouds pass by photovoltaic
cells, rapidly changing their power output.21
Solar plant operators can use energy storage,
Meeting Interconnection Challenges
static volt-ampere reactive compensators, or
other forms of reactive compensation to mitigate
The integration of DG presents new challenges
potential flicker problems.22 Distributed genera-
for distribution system planning and opera-
tion connected to the system with inverters (as
tions, principally because the configuration
are all solar PV systems) could use advanced
of power lines and protective relaying in most
inverter functionality to provide this reactive
existing distribution systems assume a uni-
compensation.
directional power flow and are designed and
operated on that assumption. Historically, the
IEEE Standard 1547 also seeks to address
penetration of DG was sufficiently small to
potential safety issues with DG, as it would
be regarded as simply a reduction in load, but
threaten the safety of utility workers were it to
this will change if DG penetrations grow. While
keep a line energized after a fault when the line
the physical wires and transformers can carry
is thought to be “dead.” 23 The standard requires
power flow in the reverse direction, DG none-
that DG units disconnect from the system
theless can have adverse impacts on system
when local faults occur or when the voltage or
reliability, power quality, and safety.17
frequency at their interconnection point falls
outside prespecified ranges. DG units are also
IEEE Standard 1547
required to detect unintentional islanding,
circumstances in which DG supplies a local
In recognition of the potential adverse impacts
portion of the grid that has been disconnected
of DG on distribution systems and the need
from the bulk power system, and disconnect
for uniform criteria and requirements for the
“within two seconds.” While the standard does
interconnection of DG, the industry collabo-
not explicitly forbid “intentional islanding,”
rated with the Institute of Electrical and
it does not specify requirements for islanded
Electronics Engineers (IEEE) to create IEEE
operation and indicates that islanding is
Standard 1547,18 first released in 2003 and later
“under consideration for future revisions”
incorporated into the Energy Policy Act of
of the standard.
2005.19 The standard’s primary intent is to ensure
that DG units do not have negative impacts on

112 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

BOX 5.1 DISTRIBUTED GENERATION’S regulation observed in this study. As an addi-
CONTRIBUTION TO VOLTAGE REGULATION tional benefit, allowing the solar PV system to
regulate the voltage at its interconnection point
In a recent study, General Electric discusses
was found to significantly reduce the need to
the potential benefits of allowing distributed
operate other voltage regulation devices
generation (DG) units to actively regulate
located along the simulated feeder. This result
voltage.24 The study simulates a 10 megawatt
suggests that allowing DG units to actively
(MW) solar photovoltaic (PV) system connected
regulate the voltage at their point of connection
to a 13.8 kilovolt (kV) feeder whose peak load
could sharply reduce voltage variation under
reached 12.3 MW. The system also had 600
high penetrations of DG. If such operation were
kilowatt (kW) solar PV inverters that could
to reduce the need for mechanical tap-chang-
simultaneously supply real power and produce
ing transformers, installed to regulated voltage,
or absorb up to 290 kV-amperes of reactive
it also would reduce maintenance costs. The
power in order to regulate voltage levels.
results likely would be similar for more moder-
Figure 5.1 illustrates the significant difference ately sized DG units on low-voltage circuits.
in feeder voltage with and without voltage

Figure 5.1 Feeder Voltage at the Point of Interconnection of a Solar PV System

1.05 1.05

1.04 1.04
Fedder Voltage (pu)
Fedder Voltage (pu)

1.03 1.03

1.02 1.02

1.01 1.01

1.00 1.00
7:00 AM 11:00 AM 3:00 PM 7:00 PM 7:00 AM 11:00 AM 3:00 PM 7:00 PM
Hour Hour

(a) Without Voltage Regulation Capability (b) With Voltage Regulation Capability

Note: The voltage scales on these plots are in a normalized measure called per-unit (pu). The normalizing constant is the
nominal voltage of the line, 13.8 kV in this case. The line is operating at approximately 1.026 pu, which is 14.2 kV.

Source: ©2010 IEEE.
Load Reprinted, with permission,
Peak Capacity from R. A. Walling and K. Clark, “Grid Support Functions Implemented in
Three 1.4 kW Load
Utility-Scale PV Three
Systems,”
3.3 kW paper presented at the Transmission and Distribution Conference and Exposition, 2010 IEEE Power
Three 1.4 kW
& Energy Society, New Orleans, LA,DesignApril Capacity
19–22, 2010. Three 3.3 kW
NRC – Maximum EIA
NRC – Probable Obama Target

Since the original creation of IEEE Standard the required capabilities of DG and necessary
1547, IEEE has supported efforts to create eight operating procedures that can be used to create
additional supplemental standards documents intentional islands,
50 thereby
50 50 partially
50 filling the
intended to extend and/or clarify the provisions previously mentioned gap in IEEE Standard
in the main standard text.25 For example, IEEE 1547. While five of these documents have been
Standard 1547.4, completed in 2011, specifies completed, three additional documents are

Chapter 5: The Impact of Distributed Generation and Electric Vehicles 113

currently in development and are expected to be Islanded Operations. IEEE Standard 1547
released over the next few years. IEEE Standard requires DG units less than 10 MVA to discon-
1547 itself was reaffirmed without change in nect when an outage (or a large voltage drop)
2008 and is next up for revision in 2013. on the main system is detected. The standard
requires disconnection in the event of
FINDING unintentional islanding and does not discuss
The potential negative impacts of DG on the requirements for intentional islanding. In the
development of IEEE Standard 1547, some
power system are being mitigated by the
argued that DG units should disconnect from
establishment of interconnection standards
the system to prevent damage to distribution
through the IEEE. system equipment and ensure the safety of
utility crews repairing outages. The require-
ment that DG units disconnect during system
Future Modifications to IEEE Standard 1547 outages effectively prevents DG units from
providing reliability benefits to surrounding
Since its initial drafting, several weaknesses customers.
in IEEE Standard 1547 have become apparent.
Grid-connected DG units, especially those The recently released IEEE Standard 1547.4
based on variable energy sources, were not as discusses the intentional use of DG to supply
prominent when IEEE Standard 1547 was first power to a disconnected part of the distribu-
created as they are expected to be over the next tion system when a fault is present in another
20 years. As the number of DG installations part of the system. Distributed generation units
grows, modifications may be needed to ensure that are connected to the grid in a way that
that the standard continues to address current complies with this standard should be capable
state-of-the-art practices and needs. This of sustaining islanded operation and providing
section discusses several changes to IEEE reliability benefits.
Standard 1547 that should be considered
if DG penetrations are to continue to grow. However, intentional islanding will require
generators that are large enough to supply
Distributed generation can complicate the regulation adequate real and reactive power to the island.
of voltage across the length of distribution feeders. It also necessitates distributed monitoring and
control systems capable of maintaining local
Voltage Regulation. Distributed generation can supply and demand balance as well as regu-
complicate the regulation of voltage across the lating the voltage and frequency within appro-
length of distribution feeders. But DG units priate ranges. These monitoring and control
connected to the grid via advanced power capabilities add cost, and owners of very small
electronics also could play a role in actively DG units are unlikely to invest in this capa-
reducing voltage flicker and regulating voltage bility. Additionally, voltage and frequency
levels at the point of interconnection (see regulation capabilities only are allowed in
Box 5.1). Power-conditioning modules within islanding operations and not when the island
DG units that are capable of voltage regulation is reconnected to the distribution system.
have improved considerably in recent years. Therefore, even though IEEE Standard 1547.4
However, IEEE Standard 1547 forbids DG units has been released, intentionally designed
from actively regulating the voltage at their islanding schemes (for example, see Box 5.2)
interconnection point. probably will be limited to larger DG units for
the immediate future.

114 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

BOX 5.2 MICROGRIDS .JDSPHSJE3%JTTUJMMJOUIFFBSMZTUBHFT0GUIF
Microgrids that are capable of separating from 160 active microgrid projects encompassing
the utility system and operating autonomously 1.2 gigawatts (GW) of installed DG worldwide,
as electrically isolated islands for extended the majority have been demonstrations and
periods of time can be formed by a part of the research pilots.28 Microgrids are expensive
distribution network incorporating distributed because they require power electronics and
generation, storage, uninterruptible power sophisticated coordination among different
supplies, or a combination of the three.26 Such customers or areas.29 It is our sense that in most
capability may be desirable for customers or situations, the cost of configuring an area as a
groups of customers that require unusually high microgrid does not justify the reliability ben-
reliability levels. Military bases, college cam- efits, which may be achieved through other
puses, hospitals, semiconductor manufacturers, means, such as backup generators. Despite the
and data centers are examples of customers challenges, microgrids have the potential to
with high reliability needs. Microgrids in island bring new control flexibility to the distribution
operation would ensure that customers within system and thus will continue to receive much
the island would still have electric power academic interest.
supplied to them despite a fault upstream.27

FINDING Distributed generation units can increase
Interconnection standards have recently current at a fault and reduce it at the protection
been revised to allow for the realization of device for the period before the DG senses the
the reliability benefits of DG by permitting fault and disconnects, making it harder to
detect a fault and complicating the coordina-
the operation of islanded distribution
tion among protection devices.30 In addition,
networks. fault currents at points of system protection
will depend on which DG units are connected
and operating at any given time. Changing fault
Active System Management currents with the introduction of DG could
lead to unreliable operation of protective
Distributed generation imposes new challenges equipment and result in faults propagating
on distribution systems that cannot be miti- beyond the first level of protection. The propa-
gated by modifying interconnection standards. gation of faults through system protection
The most prominent of these impacts is the layers can reduce system reliability and safety.
ability of DG to disrupt the operation of system
protection schemes. In contrast to the passive operation approaches
described here, new technologies promise to
Modern system protection schemes typically use allow active management of distribution
multiple layers of coordinated protection devices, systems.31 For example, it has been envisioned
including circuit breakers and fuses, to interrupt that utilities could use real-time information
current and short-circuit faults while affecting about the operation of the network and the
service to the smallest possible number of nature of connected resources to dynamically
customers. These devices are set based on fault change protective relay settings. Active manage-
current levels and other characteristics of the ment distribution system operation techniques,
local distribution network. Distribution networks such as actively using DG and loads for voltage
today are typically designed using a “fit and control and fault current level control, can
forget” approach in which settings for protection also be used to reduce the costs of mitigating
equipment remain static.

Chapter 5: The Impact of Distributed Generation and Electric Vehicles 115

challenges related to regulating voltage profiles Electric vehicles could have a disruptive
and ensuring adequate power quality with high impact on the electric grid if not
penetrations of DG.
integrated carefully.
5.2 ELECTRIC VEHICLES efficiency. HEVs, such as the Toyota Prius, have
already penetrated the automotive market.
Similarly to DG units, electric vehicles could Automotive manufacturers are now turning to
have a disruptive impact on the electric grid plug-in hybrid electric vehicles (PHEV) and
if not integrated carefully because they will battery electric vehicles (BEV). PHEVs and
connect to the distribution network to charge. BEVs have more onboard energy storage than
As we discuss below, some such vehicles will HEVs and give owners the ability to charge the
represent a larger load than a house. The extent vehicle battery from a stationary electrical
of their impact will depend on the degree and source—for example, an outlet in the garage. A
density of their penetration, charging require- PHEV contains an internal combustion engine,
ments, and the time of day they are charged. has a limited range in all-electric mode, and
uses gasoline for long trips. A BEV has an electric
“Hybrid electric vehicle” (HEV) refers to motor, but no internal combustion engine, and
a vehicle with an electric motor, an internal it has a larger battery and a range longer than
combustion engine, and limited onboard the all-electric range of a PHEV.
energy storage that improves fuel and engine

Table 5.2 Representative Electric Vehicles Available in the United States by 2012

Tesla Roadster Nissan Leaf GM Chevy Volt Toyota Plug-in Prius
Type Battery Battery Plug-in hybrid Plug-in hybrid
&MFDUSJD3BOHF 245 miles 100 miles 35 miles 15 miles
Battery Size 53 kWh 24 kWh 16 kWh 4.4 kWh
0OCPBSE$IBSHFS 9.6 kW 3.3 kW 1.44 kW 1.44 kW
Quick Charger 16.8 kW 60 kW 3.3 kW 3.3 kW
Charging Time 6 hours (onboard) 6 hours (onboard) 10 hours (onboard) 3 hours (onboard)
3.5 hours (quick) 0.5 hours (quick) 4 hours (quick) 1.5 hours (quick)
64-BVODI March 2008 December 2010 December 2010 Spring 2012
1SJDF .431
$109,000 $35,200 $40,280 $32,000

Source: Tesla Motors Inc., “Roadster Features and Specifications,” http://www.teslamotors.com/roadster/specs; Nissan Motors Company
Ltd., “Nissan Electric Leaf Car: 100% Electric. Zero Gas. Zero Tailpipe,” http://www.nissanusa.com/leaf-electric-car/; J. Wiesenfelder,
“Cars.Com Field Trial: Mobile EV Quick-Charging,” Kicking Tires, July 26, 2011, http://blogs.cars.com/kickingtires/2011/07/carscom-
field-trial-mobile-ev-quick-charging-.html; General Motors Company, “2011 Chevrolet Volt,” http://www.gm.com/content/gmcom/
home/vehicles/browseByBrand/baseball_cards/chevrolet/volt.html; General Motors Company, “Chevrolet Volt’s 240V Home Charging
Unit Priced at $490,” press release, October 6, 2010, Detroit, MI, http://gm-volt.com/2010/10/06/gm-announces-chevrolet-volt-240v-
charger-pricing-and-installation-service-provider; Toyota Motor Sales, USA, Inc., “Toyota Introduces 2012 Prius Plug-in Hybrid,” press
release, September 16, 2011, Richmond, CA, http://pressroom.toyota.com/releases/toyota+introduces+2012+prius+plug-in+hybrid.htm.

Note: “Battery Size” gives the energy storage capacity of the battery in kilowatt-hours. This parameter also provides a relative physical size
of the battery—for a given battery chemistry, e.g., lithium-ion, the battery size is directly proportional to capacity. “Onboard Charger” is
the power capability of the charger which is integral to the car. This is the rate at which the battery can be charged by the internal charger.
“Quick Charger” is the power capability of an external (optional) charger. The quick charger can provide a more rapid charge than the
internal charger, as shown by the “Charging Time” data.

116 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID

Table 5. 00047. Annual Energy Outlook 2011 (Washington. p. Department 50 2011).S. 3 (January 25. However. could be on the mance BEVs. of the expected national fleet. as illustrated by the table. Several utility. Energy Information Administration. 2010). nearly 4.32 resulting projections in Figure 5. national penetration but the regional or local penetration that is of importance to utilities.34 A projection by the U. the Toyota Prius road. ranging in energy capacity National Research Council suggests that by from about 5 kilowatt hours (kWh) for short. Transitions to Alternative Transportation Technologies—Plug-in Hybrid Electric Vehicles (Washington. Some representative EVs are shown in tially smaller penetration of EVs. By comparison.2. Chapter 5: The Impact of Distributed Generation and Electric Vehicles 117 . higher rates than PHEVs in order to limit the and government policy. one would battery costs.33 we illustrate four as a significant new distribution system load.5% range PHEVs to about 50 kWh for high-perfor. As this table illustrates. 2011). analyzed mid-range prediction from the usually lithium-ion.35 Of course.3 kWh nickel-metal hydride mation Administration (EIA) shows a substan- battery. lower than these estimates depending on Because BEVs have larger batteries. DC: National Academies Press.S. 13 million PHEVs and BEVs.PHEVs and BEVs (collectively EVs) may emerge national EV penetration.2. 2030.S.S. Varying geographic density of electric past and ongoing studies attempt to estimate vehicles will mean that some utilities or regions Figure 5. charging infra- expect that they will be charged at considerably structure. it is not the charging time.A carefully EVs as they are emerging today have batteries. competition from other vehicles. Daily Compilation of Presidential Documents 2011 DCPD No. and U. DC:50 50 50of Energy. gasoline prices. U. EVs today are penetrations could be significantly higher or being designed with a range of specifications. Energy Infor- HEV has a 1. Degree and Density of Electric Vehicle The variances in estimates of regional and local Penetrations penetration are also significant.2 Projected Electric Vehicles on the Road by 2030 50 45 PHEVs/EVs on the Road (million) 40 35 30 25 20 15 10 5 0 2010 2015 2020 2025 2030 National Research Council—Maximum U. Energy Information Administration National Research Council—Probable Obama Target (1 million by 2015) Source: Projection data from Committee on Assessment of Resource Needs for Fuel Cell and Hydrogen Technologies and National Research Council. The impact of EVs on the electrical system EVs will not initially be a concern for every depends on their market penetration.

org/resource_files/29491.3 Power Requirement of a Single Home in the San Francisco Bay Area with and without Electric Vehicle Charging 16 Peak Power Requirements (kW) 14 12 10 8 6 4 2 0 Home Home + Level I Home + Level II Charging Charging Source: Data from D. CA. 35 across a service area may under- state the challenge. BOX 5.38 2010 Oregon.5 million PHEVs and BEVs in its 40 are likely to experience penetrations much service territory by 2020. or 0. Bowermaster. “Plug-in Electric Vehicles and Their Impact: An Integrated.” presentation at Environmental Quality Policy Committee Meeting.3650For example. League of California Cities.pdf. January 21. Even within increasing concern for the local distribution regions or utility footprints.37 Even citing average penetration higher than the national average. 2011. Early integration problems are likely hotspots will need significant focus even in to arise most often when local demand rapidly the medium term. However. The narrow vertical bar represents the variance in average Figure 5. increases because of uneven distribution of vehicles. by the presence of EVs than others. EIA NRC – Probable Obama Target 50 50 50 50 thereby within utilities may be more severely impacted cluster in particular neighborhoods.3 HOUSEHOLD POWER AND ELECTRIC home peak loads for locations throughout the VEHICLE CHARGING REQUIREMENTS San Francisco area. If the current geographic distribution of HEVs incentives. as PHEVs and BEVs may 30 25 20 15 118 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID 10 . Southern California Projections of EV penetration nationally are Edison has projected a mid-case of 5% pene- 45 highly varied. Multi-Stakeholder Approach. some local regions tration. only certain system. The broad bars show the peak power requirements for the same home by Figure 5. and the deployment of the necessary is a good indicator of demand for PHEVs and infrastructure will strongly influence the BEVs. Importantly.3 compares requirements for vehicle itself and with EVs charging at the two standard charging to the average peak load of a single MFWFMTL8 -FWFM* BOEL8 -FWFM**  home near San Francisco Bay. http://www.cacities. distribution systems in California. Sacramento. and Washington likely will 2015 experience 2020 2025 2030 considerably higher penetrations than FINDING average.PGEPEVIntro(2011-01-18). geographic distribution of EVs. promotional policies.

40 However.4 kW 70 Three 3. 2011. BOX 5. January 6. This could lead to territory.39 Aggregate power requirements are also but will enable full BEV charging within minutes..EVEL )) ˆ UP TO  K7 of PHEVs in the year 2030. If 25% of the national fleet At the residential level.3 kW Load 40 Peak Capacity Three 1.3 kW 30 50 Design Capacity 40 kW kW 20 30 20 10 10 0 0 12:00 AM 6:00 AM 12:00 PM 6:00 PM 12:00 AM 6:00 AM 12:00 AM 6:00 AM 12:00 PM 6:00 PM 12:00 AM 6:00 AM (a) 25 kVA Transformer (b) 50 kVA Transformer Source: J. this increase Charging BEVs will have more impact on the is unlikely to materialize as there will be distribution system due to their higher-power temporal diversity in the time of arrival at charging and higher energy capacity than home and most charging will probably be at Chapter 5: The Impact of Distributed Generation and Electric Vehicles 119 .4 kW 60 Design Capacity Three 3. a 50 kVA transformer of three PHEVs Figure 5.4 THE EFFECT OF UNCOORDINATED charging at two rates. tion. On the other hand. and the 25 kVA impact of plug-in hybrid electric vehicles (PHEV) transformer exceeds even its peak (short-term) on distribution system components in its service capacity rating in both cases. the power requirements could be are expected to charge using Level I chargers. “Plug-in Electric Vehicle Overview. on a warm summer day. two charging power levels: 50 50 50 50 A study in 2008 estimated that if each North s . each area would require less than a 5. DTE Energy. unlikely to require significant upgrades to the bulk power system. 1. Figure 5.4 shows the impact on both voltage dips. PHEVs are expected NRC – Maximum EIA to comprise a majority of EVs. up to 30% of generation capacity if simultane- while BEVs are expected to charge at Level II. webcast hosted by Intelligent Utility. service interruption.EVEL ) ˆ UP TO  K7 American Electric Reliability Corporation (NERC) region were to have a 25% penetration s .” presentation at A Tale of Three Cities.3 kW. Therefore the potential Engineers has established charging standards impact of these two types of vehicles likely (in Standard J1772) that cover the following will be similar.5% increase in genera- Level III has not yet been standardized in the U.4 kilowatts (kW) CHARGING ON TRANSFORMERS: and 3. and a few PHEVs EVs are expectedNRC – Probable to charge at one of three Obama Target charging simultaneously could have an impact power levels.S. LeBrun.4 Impact of Three PHEVs on Transformer Loading 25 kVA Transformer with Uncontrolled PHEV Load** 50 kVA Transformer with Uncontrolled PHEV Load* 50 Load 80 Peak Capacity Three 1. Electrical Vehicle Charging PHEVs. the majority of PHEVs were PHEVs. The Society of Automotive similar to one BEV. ously charged at 6 kW. and trans- a 25 kilovolt-ampere (kVA) transformer and former failure. These transformers are loaded beyond their design DTE Energy recently conducted a study on the capacity for both charging rates.

4).” incentives for V2G operation also appear weak.XBSSBOUZJTTVFT DPNQMFYJUZBOE also have a much more limited impact on expense of added controls and communication vehicle battery life. tional) charging rate—for example. communication between vehicles and the V2G concepts also face other substantial utility. even a few lifetimes due to temperature-induced insula- EVs on a distribution feeder could overload tion aging from capacity overload. higher charger power ratings. While still requiring unidirectional vehicle chargers and controls. it has been envisioned that vehicles could participate in frequency regula.45 This concept is most Beyond the technical challenges. Institute (EPRI) project suggest that peak nisms to influence the timing of vehicle charging. The partici- batteries could provide various types of operat. 120 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . reducing peak loading load conditions.42 EVs would be a substantial but not dominant expense that most likely all system ratepayers Influencing Electric Vehicle Charging would bear.43 Early is some concern that EV charging could impact results from an ongoing Electric Power Research local distribution systems. in many cases at the same time as neighborhood Influencing the timing of vehicle charging can load peaks. discussed in the next section. forcing utilities to invest in and improving load factor—that is. most (see Chapter 8) in neighborhoods where EV owners will plug in their vehicles and begin EVs required system upgrades. energy stored in vehicles' has historically been relatively low. the vehicles providing frequency regulation price paid to participants for regulation services services. decreasing V2G operations would require substantial and their rate to provide up regulation. While most often discussed in the context of In those markets with a regulation product. the charger requirements would be technical challenges—degradation of battery much simplified. cause the prices to decline further. sale markets.5 VEHICLE-TO-GRID OPERATIONS sufficiently charged for driving the car—and are Some observers have suggested that the flow unlikely to achieve widespread deployment in of energy between the power system and EVs the short term. and charging. and the relatively small amount use for this purpose of commercial EV fleets of energy involved if the battery is always to be which have deterministic charging patterns. the ratio BOX 5. though costs also could be recov- ered through higher fixed capacity charges If electric rates do not vary over time. charging when they arrive home each day. This would exacerbate local peak avoid these outcomes. This mode of operation would MJGF 0&.44 The that feeder and associated transformers additional cost to the system of provisions for (see Box 5. pation of EVs in these markets would likely ing reserves. could be bidirectional. Because increases in the number of EVs on a trans- an EV charging with a Level II charger is a former could yield decreases in transformer bigger load than the average house. the economic often referred to as “vehicle-to-grid operation. levels lower than 6 kW.41 Nevertheless. requiring mecha. Particularly attractive is the with the utility. there expanded infrastructure (see Box 5.5). in theory. A more cost-effective alternative use of EVs tion or other reserve markets by supplying would be to provide regulation or operating reserves only through control of their (unidirec- energy to the grid. and increas- expensive modifications to conventional ing it for down regulation. In regions with organized whole.

a two-way enced by policy or controls to produce a flat communication link between the charging load between 6 p. pricing would improve system-wide load factor. or they can be signaled a day To give an idea of the effect.1% load probably would develop at the time of and 31.9% time-of-day. the regional station and utility. a secondary peak vehicles on arrival at home. According The requirements include a tool to control to a 2002 study.between average and peak power. be inversely proportional to the historical load but ensured that enough energy was transferred A weakness of price signals for the distribution to the vehicle for its next day’s trips. but are unlikely to have a substantial impact on Customer reactions to such a control scheme feeder overloading. voltage regulation (see Box 5.m. penetration level from a low of 15% in and the state of charge of each of those vehicles. charging power at each vehicle. but advanced metering differences between PHEVs and BEVs must be infrastructure would help enable such a scheme. an hour ahead.48 California to a high of 73% in Texas. Implementing it might require a price break for bution circuits by remotely controlling charging.4% of the local distribution transformers price change.5). participating customers and a mechanism for The exact mechanisms for accomplishing this overriding the direct control at some cost. even this high penetration neighborhood congestion levels. Utilities can more directly would be complicated and potentially negative. The effect by staggering the rate structure controlled charging case set charging rates to geographically. a variety of ways: they can be static and based on time of use.49 They assumed two charging rates set to avoid the most expensive times of day. period-ahead. in ways acceptable to consumers have not yet In designing measures to influence demand. and even real-time of transformers.46 Controlling the charging of EVs may enable them to benefit utility operations by providing FINDING ancillary services such as frequency and/or For real-time pricing to be effective. EVs must be capable of automatically FINDING responding to price signals. With system is that they offer little insight into controlled charging. In the uncon- If the time-of-use tariff were to have uniform trolled case.47 Regulators might counter this at the 3 kW and 10 kW rates. and knowledge of the system generation capacity could handle a PHEV state. the number of vehicles requiring charging. if vehicle charging were influ. overloading 22. for the vehicles: 3 kW and 10 kW. owners would begin charging their timing over the whole system. the been fully worked out. respectively.m. mitigate the impacts of EV charging on distri. Using time-differentiated tariffs or central Two viable methods for influencing the timing control schemes to discourage electric of charging have emerged: time-differentiated vehicle charging at peak times can improve tariffs and centralized charging control structures. and 6 a. system operation and avoid requiring Time-differentiated tariffs can be structured in capital investments in new infrastructure. As a result. Vehicle lated controlled and uncontrolled charging of owners might respond to time-of-use pricing an aggressive 75% penetration of EVs in the by simply putting the vehicle charger on a timer Netherlands. researchers simu- ahead.. With their fuel-powered Chapter 5: The Impact of Distributed Generation and Electric Vehicles 121 . of vehicles would require upgrading only 1. carefully considered. or in real time.

3 million to 40 million by 2030. market. PHEVs and BEVs. BEV owners may want to keep their batteries fully charged. RECOMMENDATIONS Plug-in hybrid and battery-powered electric Distributed generation and plug-in electric vehicles. The number of types of generation and loads that have been EVs on the road in the U. range and operational flexibility. consumer preference and income. 122 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . Ultimately. advantages of cogeneration and policies that charging infrastructure availability. In particular. have vehicles are qualitatively different from the begun to enter the U. engines. IEEE Standard are expected to cluster in select high-income 1547 was a first attempt at establishing uniform and eco-conscious neighborhoods. Between the interconnection criteria for small generators two types of EVs. permit voltage regulation by distributed generators to enable them to help maintain The introduction of time-differentiated tariffs is not only being considered in the context of distribution voltages within limits. As discussed in Chapter 7. deployment of these technologies could reduce the total costs of integrating high 5. and therefore may resist methods that DG interconnection standards should influence vehicle charging. and enabling DG units to actively to benefit utility operations by providing ancillary regulate the voltage at their interconnection services such as frequency and/or voltage regulation. flexibility in the timing of their vehicle modifications to this standard will become charging. important than the magnitude of penetration is the fact that it is not expected to be uniform Growth in DG will arise from the economic across the nation because of state incentives. Widespread deployment of these tech. more important. points would ease the burden of providing uniform and constant voltage along distribu- take advantage of lower rates. by 2030 will depend connected to the electric power system in the on a number of factors that are difficult to past. By comparison. More distribution system. adding provisions for islanded operation of DG units Controlling the charging of EVs may enable them would permit them to enhance the reliability of supply. as DG penetrations continue to grow.S. In fact. collectively referred to as EVs. demand response programs aim to shift total demand away from peak periods to realize both Growth in DG also will motivate the deploy- short-run operational benefits and long-run ment of active distribution system management investment efficiency improvements. and owners may be more willing to increasingly important. National projections range from as few nologies will change the requirements of the as 3.S. including the deployment of vehicles could make these programs even additional communication and sensors. EVs such as rooftop solar panels. predict. charging when they have a R E CO M M E N D AT I O N chance. PHEVs offer their owners more However. PHEVs will achieve greater and included a range of provisions to mitigate penetration than BEVs due to their superior many of the challenges associated with DG. Electric technologies. and encourage distributed renewable generation.3 CONCLUSIONS AND penetrations of DGs. tion feeders.

Otherwise. If regulators and utilities appropri. ately influence charging so that it mostly does not coincide with the system peak demand. EVs will improve system load factor and will not cause unmanageable disruption to the bulk generation and transmission system. Chapter 5: The Impact of Distributed Generation and Electric Vehicles 123 .The degree to which EVs pose a stress to the R E CO M M E N D AT I O N power grid depends on their local penetration Utilities in regions with potentially high rate. mechanisms to incent off-peak charging. integrating these loads will require more investment in equipment. as well as the power and time at which penetrations of EVs should explore they charge.

January 28–February 1. Grenard. NY: Interstate Renewable Energy Council. Systems. Barker and De Mello. “Smart Distribution: Coupled Systems” presented at the IEEE PES Transmission Microgrids. “Shape of the Microgrid. 8 U. Department of Energy.org/.com/ to Quantify the Technical Benefits of Distributed facts-and-figures/retail-price-environment/ Generation. Barbose. 23 and Demonstration Projects. Ramakumar. gov/~webworks/cppr/y2002/rpt/112701. 2002). Energy Policy Act of 2005. Assessment to Enhance the Business Case for the and Society (Oak Ridge. 1547-2003 2010).S. and C. http://www. Mamo. An Overview of Ongoing Research.ieee. IEEE Standards Association. Dispersed Generation. P. IEEESTD. Marnay.S. DC.ornl. of Distributed Generation and Rate-Related Issues that May Impede Their Expansion: A Study Pursuant U. Canada. Clark. Coste. 2005).1109/ 2010/07/IREC-Solar-Market-Trends-Report. see note 8 above. WI: University of Laboratory. New Orleans..” IEEE Transactions on Energy solar-electricity-prices. no. 2005.S. 15 The NPD Group. N. LBNL-50829 Generation (Fairfax. and Department of Energy. Reaffirmed 2008). Connecting Distributed Energy Columbus.S. 2007).pdf. on Environment and Electrical Engineering. DC: Strategy” presented at the IEEE Power & Energy U. 2001. 1074–1082. Resources to the Grid: Their Benefits to the DER R. 2009. Solar Market Trends 2009 for Interconnecting Distributed Resources With (Latham. Stovall.” presented at IEEE Power Engineering Society Summer Meeting. H. Wisconsin–Madison. S.” IEEE Power & Energy Magazine 5. G. and C. Conversion 19 (2004): 764–773. the Utility. Other Customers. and A.” IEEE Std. see note 6 above. J. and Expansion in Southern California Edison Service Energy Storage (SCC21). 2010. LBNL-1516E (Berkeley. “Distributed Generation and Power 22 to Section 1817 of the Energy Policy Act of 2005 Quality” presented at the International Conference (Washington. Peterman. “Microgrids: Karpacz.S. “Distribution Automation: The 3 U. 20 Lawrence Berkeley National Laboratory. Albuyeh. “Determining 21 7 Database of State Incentives for Renewables the Impact of Distributed Generation on Power & Efficiency. August 8. 970. Energy Information Administration. TN: Oak Ridge National CERTS Microgrid (Madison. REFERENCES Wiser. from 1998-2007. S. and C.S. 11 U. Lasseter et al. Exploring Distributed IEEE Standards Coordinating Committee on Fuel 25 Energy Alternatives to Electrical Distribution Grid Cells. 4 Ibid. 2009. Inc. “An Approach Research and Analysis. the Sun: The Installed Cost of Photovoltaics in the U. Robio. Photovoltaics. W. 19 6 R. VA: Energy and Environ. July 16–20.94285. Laboratory. A. Barker and R. Energy Information Administration. U. Ipakchi and F. CA: Lawrence Berkeley National mental Analysis. http://www.. Eto.” Proceedings of IEEE 99. 2010). “Grid of the Future.S. “IEEE Standard 18 5 L. no. 14 W. July 26–30. Value and Technology 27 Owner/Customer. 2 (2009): Power Annual 2009 (Washington. May 10–13.org/wp-content/uploads/ (Issued 2003. http://irecusa. Mallet. 124 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . Department of Energy. “Solarbuzz: Solar Market 16 1 P. Annual Cornerstone for Smart Grid Development Electric Power Industry Report (Washington.” http://solarbuzz. CA: IEEE Standards Association. Environmental Benefits of Distributed The CERTS MicroGrid Concept. Department of Energy. Electric Power Systems. 2000). P. 2011). Chiradeja and R. 52–62. Iravani. Poore et al. Sherwood.pdf. Calgary. “Grid Support 24 Functions Implemented in Utility-Scale PV 9 R. R. Tracking 109-58. Poland. R. Walling and K. 2009). see note 18 above. Laboratory. WA.H. T. Bluestein. 10 T. Kingston and T. The Potential Benefits 2000.S. 2010_7-27-10_web1.2003. Society General Meeting. 1254. P.” presented 13 U. see note 21 above. Asano. no.dsireusa. Electric IEEE Power & Energy Magazine 7. Barbose. US Congress. DC: U.S. Wiser. see note 8 above. 119 STAT. Lasseter and J. April 19–22.” http://grouper. 2002). Department of Energy.. Sec.” 17 2 U. doi:10. TN: Oak Ridge National Interconnection Standards. L. Development. F. and Peterman. and N Hatziargyriou. OH. Khan. (Berkeley. org/groups/scc21/dr_shared/. Marnay. Siddiqui. Pub. “1547 Series of Territory (Oak Ridge. S. at IEEE Power Engineering Society Winter Meeting. A. and X. 6 (2011): and Distribution Conference and Exposition. LA. 4 (2007): 78–94. Integration of Distributed Energy: 26 12 J. De Mello. 2010). Seattle. R. Lasseter.

com/hybrid-market-dashboard/ N. November 9. C. Department of Energy. National IEEE Spectrum 47. CA. CA: Electric Power Car. 2011. F. Prague. “Cities Prepare for Life with the Electric Gas Emissions (Palo Alto. J. Market for PHEVs: M. Potential Policy Committee Meeting.S. 39 Transportation and Achieving Energy Security Hadley and Tsetkova. Vehicles and Impacts to the US Power System: UCD-ITS-RR-08-22 (Davis. Eto. WA: Pacific Northwest National Davis. 1 (2010): 13–14. University of California. Germany: Ronald Berger Strategy Consultants. Kurani. 2010. J. “Integrating Plug-in-Electric 44 oiaf/archive/ieo09/transportation. 2008).gov/ A. Gerkensmeyer. Republic.html. Electricity Distribution Networks: Summary of the D. “Dispersed november-2008-dashboard-25328. Committee on Assessment of EV Conference. http://www. International Energy Outlook 2009 above. and F. Woody and Electric Vehicles Volume 1: Nationwide Greenhouse C. Director of Plug-in Electric Vehicle 37 DG-GRID Project Results ECN-E--07-083 (DG.S.com/hybrid- M. Bernhart. DC: Pike Research. no. Deutsche Bank Securities Inc. Powertrain Minneapolis. Book et al. Approach. see note 33 above. Maitra et al. Energy Information Administration. MA: Boston Consulting Group. W. 2008). Group. July 25–29. 2009. ibid. Electric Cars: Required Solutions” presented at The Networked Plugged In (2008). W. and Electrification Coalition.doe.. (1999): 22–28. 40 (Washington. J.. (Richland. Alternative Fuels & 38 Resources Defense Council. Lasseter and J. Regulatory Improvements for 31 sales-dashboard/december-2008-dashboard- Effective Integration of Distributed Generation into focus-production-numbers-25416. 2007).org/resource_files/29491. The Comeback of the Electric Car? 33 Vehicle Distribution System Impacts” presented at (Boston. and Generation Impact on Distribution Networks. The Early U. Readiness.cacities. 2009). 2007).Valentine-Urbschat and W. 2009. Kim. and P. on Electricity Distribution. and T. December 10. Energy Information Laboratory. “Plug-in Electric Vehicles and 43 Technologies—Plug-in Hybrid Electric Vehicles Their Impact: An Integrated. Research Institute. Electric Power Research Institute. National 32 U. Department of Energy.eia. Technologies. see note 33 above.html. Ridge National Laboratory. 2 Production Numbers. 29 36 “November 2008 Dashboard: Sales Go From Bad to Worse. Czech WA: Pacific Northwest National Laboratory. K. Taylor et al. 2010. and P. see note 41 Administration.. DC: The National Academies Press. (Washington.” HybridCars. Dasso.afdc. Multi-Stakeholder (Washington. DeSteese. http:// Regional Power Generation (Oak Ridge. February 15. CA. PNNL-19165 Transportation Studies. Sacramento. “Evaluations of Plug-in Electric 41 M. 2010). and Taylor et al.. Krauss. M. Transitions to Alternative Transportation D. Recharge Technical Challenges of Plug-in Hybrid Electric Potential. U.” The New York Times. 2020—The Future Drives Electric (Munich. TN: Oak www. March 22. S. January 21. “Speed Committee on Assessment of Resource Needs for 34 Bumps Ahead for Electric-Vehicle Charging. Scheepers et al.” “December 2008 Dashboard: The Key Is IEEE Computer Applications in Power 12. 35 Microgrid Deployment Tracker” (Boulder. CO: Annual Energy Outlook 2011 (Washington. PNNL-17441 (Richland. 2008. Environmental Assessment of Plug-in Hybrid energy. see note 21 above.gov/afdc/laws/2010. conversa- GRID Consortium.com. Canard. January 13. San Francisco. K. 2009).” HybridCars. Asmus and B. Vehicles with the Distribution System” presented Balducci. June 8–11.html. Dumas.S. G. “Executive Summary: 28 U. Hadley and A. Fairley. and J. http://hybridcars. and 30 http://hybridcars. Davis.. Electrification Roadmap: Revolutionizing Hadley and Tsvetkova. Design Priorities and Energy Impacts.S. “PEV Adoption at Scale: Grid Challenges. Southern California Edison. Bowermaster. 2011. 2009). P. IEEE Power and Energy Society General Meeting. 42 2009). 2011). http://www. Research Council. Hadjsaid. C.com. National Research Council. Anticipating Consumer Awareness. Plug-in Hybrid Electric Vehicle Market at 20th International Conference and Exhibition Penetration Scenarios. CA: Institute of Distribution System Analysis. and Charles Clark Advanced Vehicles Data Center. MN. J. Barker and De Mello. see note 27 above. tion with the authors.. League of California Impacts of Plug-in Hybrid Electric Vehicles on Cities. Tsvetkova. DC.pdf. R.” Fuel Cell and Hydrogen Technologies. Kintner-Meyer. 2010. Axsen and PGEPEVIntro(2011-01-18). no. U. Resource Needs for Fuel Cell and Hydrogen 2010. 2008).S.” presentation at Environmental Quality 2010). DC. Chapter 5: The Impact of Distributed Generation and Electric Vehicles 125 .

Residential Low Voltage Networks” presented at intelligentutility. J. Tomi´c. Kempton and J. 126 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . 45 S. “Vehicle-to-Grid Power Fundamentals: Calculating Capacity and Net Revenue. Carson.” Intelligent Utility. http://www. Kintner-Meyer. P.intelligentutility. K. F.” Journal of Power Sources 195. Verzijlbergh. and W.. Pratt. http://www. 2007). 47 Economics of using Plug-in Hybrid Electric Vehicle Taylor et al. 8 (2010): 2377-2384. Impacts Assessment of Plug-in Hybrid Vehicles on Electric Utilities and Regional U. the Netherlands. Whitacre.S. B. 2011. no. April 11–13. see note 41 above. 2010. P. February 17. no. “The Impact of Controlled EV Charging on Intelligent Utility. 48 Battery Packs for Grid Storage. Carson. 46 M. W. v2g-question. D.com/ article/10/02/v2g-we-ve-got-whole-package. Power Grids: Part 1: Technical Analysis (Richland.. Delft. “V2G: We’ve Got the Sensing and Control. A. February 24. Slootweg. Peterson. M. and J. 1 (2005): 268-279. 49 “To V2G. 2010. see note 41 above. Z. C. Lukszo. “The Taylor et al.com/article/10/02/v2g-or-not. Schneider and R. IEEE International Conference on Networking. J. Whole Package. R. G. Apt. P. G. Or Not to V2G? That is the Question!” Ilic. WA: Pacific Northwest National Laboratory.” Journal of Power Sources 144.

reliability. Section 6. and restoration systems as well as voltage and power flow optimization systems. and most of this loss occurs in the s REGULATES VOLTAGES WITHIN A LIMITED RANGE1 distribution system. We describe the potential of more fully integrated distribution management systems. is lost between generators and customers.3 discusses the importance of learning from technology pilot programs and early deployments. We find that many individual technologies that could enhance the distribution system are complementary.Chapter 6: Enhancing the Distribution System In this chapter.1 introduces technologies that promise to enhance distribution system operation. isolation. and substations and consumers. Detailed information generated from pilots and early deployments will enable utilities and their regulators to make more informed decisions on both investment priorities and system design. We describe several categories of operational benefits of AMI and analyze the costs and benefits of AMI based on recent utility regulatory filings. and the benefits from their deployment will be greatest when utilities use an integrated approach to system modernization. In addition to transporting electricity. Section 6. and power quality of distribution systems. These include automated fault detection.2 2ELIABILITY FOR INDIVIDUAL CUSTOMERS VARIES WIDELY. the distribution system s RESTORES SERVICE FOLLOWING INTERRUPTIONS s TRANSFORMS VOLTAGE TO THE APPROPRIATE LEVEL Today. and automation and information technologies promise to improve the efficiency. New sensors. Section 6. management systems. we focus on the potential for new technologies to enhance the performance of distribution systems. The distribution system is the portion of the s DETECTS FAULTS AND OTHER ABNORMAL SITUATIONS electric power system that carries power the on the distribution system and takes action to few miles remaining between transmission protect people and system components.4 describes our conclusions and details this chapter’s one recommendation. communication equipment.2 contains a more in-depth discussion of one particular distribution system technology: advanced metering infrastructure (AMI). in the U. to be discussed in more detail in Chapter 7. The costs and benefits of new distribution technologies are subject to significant uncertainty. less than 7% of total electrical energy for customer use. we recommend that policy makers work to ensure that comprehensive data from Recovery Act–funded programs are shared as widely as possible throughout the industry. We highlight the importance of the nonoperational benefits associated with AMI. Recognizing the importance of data from distribution technology pilot programs and early deployments. Section 6. Finally.S. then introduce several new distribution system operation applications.

BUT AVERAGES IN Chapter 6: Enhancing the Distribution System 127 .

S.5 and 2 power INVESTMENT SHOULD BE THAT THERE IS A TRADE OFF interruptions per year and between 2 and 8 BETWEEN COST AND LEVEL OF SERVICE hours without power. tell us that principle in any discussion of new distribution customers can expect between 1. different regions of the U.3 According to the rele- VANT LITERATURE.

but little attention has been paid to measurement. and publica- 80% of interruptions are due to problems in the tion of standardized performance metrics distribution system. collection.4 The to this assessment. differences in definitions and data collection able to the distribution system can be much processes. ACROSS 53 DISTRIBUTION UTILITIES 0REVIOUS ATTEMPTS TO COMPILE SUCH DATA HAVE HIGHLIGHTED percentage of customer outage time attribut. -ORE FUNDAMENTALLY.  OF INTERRUPTIONS ARE DUE -EASURING SERVICE QUALITY IS AN IMPORTANT INPUT to problems in the distribution system.

AS OF .

the only 35 state public utility commissions distribution system was responsible for at INCLUDING THE $ISTRICT OF #OLUMBIA REQUIRED least 95% of interruptionsi in the Southern reporting of standard distribution performance #ALIFORNIA %DISON SERVICE REGION FROM n metrics. though this represented an increase  AND  OF &LORIDA 0OWER  . higher in some specific cases.IGHT OVER A PREVIOUS COUNT IN 7 Whereas INTERRUPTIONS IN 5 %UROPEAN REGULATORS HAVE MADE A COORDINATED EFFORT TO REGULARLY COMPILE COMPREHENSIVE DATA 0OWER QUALITY REFERS TO THE DEGREE TO WHICH from distribution utilities across Europe and VOLTAGE LEVELS ARE ACCURATE AND CONSTANT. for example.

TELEVISIONS. as discussed in Chapter 5.8 no effort on that scale has taken QUALITY CAN BE COMPROMISED BY CERTAIN TYPES PLACE IN THE 53 )MPROVING THESE DATA COULD of electrical loads as well as distribution system FACILITATE EVALUATION OF INVESTMENT OPPORTUNITIES EQUIPMENT $ISTRIBUTED GENERATORS MAY going forward. 0OWER QUALITY ISSUES GENERALLY DO NOT CAUSE 6. AND account for definitional and geographical THE VOLTAGE WAVEFORM IS A PURE SINUSOID 0OWER differences. contribute as well.1 OPPORTUNITIES IN DISTRIBUTION PROBLEMS WITH CUSTOMER EQUIPMENT BECAUSE SYSTEM OPERATION modern power supplies for phones. as we recommend in Chapter 8. computers.

and ADVANCED INFORMATION TECHNOLOGY WILL ENABLE 0ROVIDING EFlCIENCY. sensor technologies. AND OTHER ELECTRONICS CAN ACCEPT A The integration of new communications VERY LARGE VARIATION IN VOLTAGE LEVEL AND SHAPE infrastructures.

RELIABILITY.

AND POWER NEW DISTRIBUTION SYSTEM CAPABILITIES OVER THE QUALITY IS EXPENSIVE FOR UTILITIES AND RATEPAYERS.

NEXT SEVERAL DECADES 7HILE MANY OF THESE BUT INEFlCIENCY AND POOR RELIABILITY ALSO HAVE TECHNOLOGIES HAVE EXISTED FOR SOME TIME.

THEIR COSTS 5P TO A CERTAIN POINT.

distribution systems has system hardware represent a net gain to society. INVESTMENTS IN deployment in U. been limited. "EYOND THAT LEVEL.S.

i 0ERCENT OF INTERRUPTIONS ARE CALCULATED USING 3!)$) SYSTEM AVERAGE INTERRUPTION DURATION INDEX VALUES 3!)$). ADDITIONAL IMPROVEMENTS FROM INVESTMENTS MAY NOT BE COST EFFECTIVE 4HE 5TILITIES HISTORICALLY HAVE EMPLOYED A VARIETY OPTIMAL LEVEL OF SERVICE VARIES GREATLY DEPENDING of manual or semi-automated systems to ON INDIVIDUAL CUSTOMER NEEDS !N UNDERLYING monitor system status. manage work crews.

A RELIABILITY INDICATOR COMMONLY USED BY ELECTRIC UTILITIES.

IS THE AVERAGE OUTAGE DURATION FOR EACH CUSTOMER SERVED.

USUALLY MEASURED IN MINUTES PER YEAR 3EE THE GLOSSARY FOR DETAILS 128 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .

BE DISCOVERED WHEN CUSTOMERS CALLED THE UTILITY track problems.detect outages. manage infra. a distribution system failure would utilities still use paper maps to keep records. Then the contrast. In to complain about lack of power. electronic maps allow for easy utility would send out a crew to track down UPDATING. to outdated. and manage work crews. Surprisingly. manage assets. $-3S IN USE TODAY RUN THE GAMUT FROM MODERN structure upgrades. and perform other tasks. some distribution Historically.

Electronic maps are only this history. $ISTRIBUTION SYSTEMS TODAY ARE A PRODUCT OF with work crews. today’s distribution one example of the changes that upgrades companies still employ essentially the same IN DISTRIBUTION SOFTWARE CAN BRING (OWEVER. QUICK SEARCHES THROUGH LARGE and repair the problem. and standardization was largely nonexistent. and clear communication DEVELOPED AD HOC MANAGEMENT SYSTEMS TO TRACK AND ASSIGN LINE CREWS .ACK OF REAL TIME The integration of new communications instrumentation throughout distribution infrastructures. system capabilities over the next several decades. and advanced NETWORKS OFTEN LIMITED THE EFFECTIVENESS OF information technology will enable new distribution these tools. Some utility companies geographic areas. sensor technologies. In many areas.

IN process for locating and managing outages an industry that prizes reliability and depends AS DECADES AGO )N OTHER AREAS.

SYSTEMS HAVE on tested operating procedures to maintain EVOLVEDˆBUT NOT ALWAYS IN THE SAME DIRECTION.

This heterogeneity reflects period of heightened risk as personnel become the general lack of standardization as well as accustomed to new practices. such fundamental changes come at resulting in heterogeneous distribution systems THE COST OF EXTENSIVE TRAINING PROGRAMS AND A around the country. 4HE TIME. reliability. differences in geography and regulatory prac- tices across state boundaries.

MONEY.

CUSTOMIZATION IS NECESSARY ! VARIETY OF EQUIP- ment of significantly more sophisticated ment throughout the distribution system can DISTRIBUTION MANAGEMENT SYSTEMS $-3 THAT BE ENABLED TO INTERACT WITH $-3S -ANY NEW CAN HANDLE AND FULLY INTEGRATE A WIDE VARIETY OF SENSORS. AND EXPERTISE REQUIRED TO IMPLEMENT A NEW $-3 CAN VARY GREATLY 2ECENT ADVANCES IN INFORMATION AND COMMUNI. BETWEEN UTILITIES BECAUSE A SIGNIlCANT LEVEL OF CATION TECHNOLOGIES HAVE ENABLED THE DEVELOP.

PROTECTION EQUIPMENT.

The number of sensors that a throughout the distribution system. UTILITY DEPLOYS AND INTEGRATES WITH ITS $-3 THE ART $-3S OFTEN CAN SIMULATE DISTRIBUTION IMPACTS THE COST AND TIME REQUIRED FOR IMPLE- system power flows in near-real time. AND CONTROL SYSTEM MANAGEMENT TASKS "EYOND PROVIDING DEVICES HAVE COMMUNICATION AND COMPUTATION REAL TIME VISIBILITY INTO THE STATE OF ASSETS capabilities. state-of. helping mentation and the usefulness of the resulting operators anticipate or respond to potential $-3 4O UPGRADE ITS $-3 FROM A SOFTWARE PROBLEMS &OR EXAMPLE.

THESE SYSTEMS CAN ALLOW SYSTEM DEVELOPED IN HOUSE TO A STATE OF THE ART operators to determine whether changing the VENDOR SOLUTION.

ONE #ALIFORNIA UTILITY ATTRIB- configuration of the distribution network in UTED  OF THE ESTIMATED IMPLEMENTATION COST response to faults will result in a system with TO LABOR.

)4.

Chapter 6: Enhancing the Distribution System 129 . MODERN $-3S PROMISE TO MAKE DISTRIBUTION SYSTEM OPERATORS MORE AGILE AND RESPONSIVE to real-time system conditions. AND EQUIPMENT UPGRADES AND  APPROPRIATE VOLTAGE LEVELS AND CURRENT mOWS OF THE COST TO SOFTWARE VENDOR SERVICES9 In conjunction with other technologies.

sensor. TION SYSTEM OPERATORS USE 3OMETIMES VIEWED and advanced information technologies AS A COMPONENT OF $-3S. The use of new communication. /UTAGE MANAGEMENT SYSTEMS /-3 ARE FINDING another tool in the software suites that distribu.

/-3S TYPICALLY USE can result in the integration and automation system models and graphical user interfaces of many distribution system functions. locate outages. State-of. and manage repair crews. to handle customer complaint calls. information systems. yielding reduced costs and improved the-art OMSs can incorporate geographic reliability. and REAL TIME DATA FROM SENSORS TO PROVIDE SOPHIS TICATED. electrical models.

Isolation.ABORATORY STUDY QUANTIlES THE ECONOMIC action. REAL TIME VISUALIZATION OF THE SYSTEM Automated Fault Detection. consider an automobile accident that benefit of reducing outage time: using data knocks down a pole supporting distribution FROM CUSTOMER SURVEYS.ATIONAL AUTOMATIC &$)2 !S AN EXAMPLE OF &$)2 IN .AWRENCE "ERKELEY . DURING OUTAGES )MPROVED WORK MANAGEMENT and Restoration software and real-time outage information with clearer communication between mobile units -ORE EXTENSIVE DEPLOYMENT OF CIRCUIT BREAKERS and control centers can increase the speed with AND COMMUNICATION DEVICES COMBINED WITH which crews reach faults and the safety of their ADVANCED CONTROL ALGORITHMS CAN ENABLE ACTIONS ! RECENT .

ITS AUTHORS CONCLUDE WIRES.

MATIC &$)2 IN PLACE THAT ALSO INCLUDES MULTIPLE  FOR SMALL COMMERCIAL AND INDUSTRIAL FEEDERS AND SECTIONALIZING SWITCHESˆSWITCHES CUSTOMERS. INTERRUPTING SERVICE TO CUSTOMERS that the cost of an outage lasting one hour is A distribution system with software for auto- approximately $4 for residential customers.

AND .

11 Of ISOLATE THE SITE OF THE FAULT THE DOWNED WIRES . FOR MEDIUM AND LARGE THAT DIVIDE LINES INTO INDEPENDENT SECTIONSˆCAN commercial and industrial customers.

COURSE.

and COMPUTING AVERAGE VALUES REQUIRES A RANGE OF PROVIDE A SECONDARY PATH FOR SERVICE TO CUSTOMERS assumptions regarding outage and customer who would otherwise be without power. This self-healing capability has the potential to Integrating communications. characteristics. IMPROVE RELIABILITY. THE ACTUAL COSTS OF OUTAGES VARY WIDELY perform analysis to determine the extent of across different customers and regions. IT infrastructure. and damage and options for reconfiguration.

ENHANCE CUSTOMER SERVICE.

AND SENSORS WITH $-3S WILL ENABLE A RANGE OF and reduce operation and management costs. $EPLOYING AUTOMATED &$)2 IN LEGACY SYSTEMS Two prominent examples are automated fault CAN SPEED THESE PROCESSES BY REMOVING THE NEED DETECTION.12 new distribution system operation applications.

ISOLATION.

AND RESTORATION &$)2.

In KNOWN AS SELF HEALING AND OPTIMIZATION OF distribution systems that already use automated SYSTEM VOLTAGES AND POWER mOWS 4HE TERM &$)2 PROCESSES. ALSO for human action and decision-making.

and reduce the time of outages and the number of operation of the distribution network. We CUSTOMERS AFFECTED &OR EXAMPLE. THE ADDITION OF ADVANCED “distribution automation” is often used to control algorithms and more finely sectional- generically describe these applications of new ized distribution circuits in these systems can technology to the maintenance. control.

/KLAHOMA address these new applications in the rest of 'AS  %LECTRIC #OMPANY REDUCED ITS OUTAGE this section.13 130 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . TIME BY BETWEEN  AND  ON THREE CIRCUITS AFTER INSTALLING HARDWARE ENABLING &$)2 capability.

Oklahoma Gas & Electric Company -ANUFACTURERS DESIGN EQUIPMENT

SUCH AS
reduced its outage time by between motors, to operate optimally within the range
OF Ò  OF NOMINAL VOLTAGE 3UPPLY VOLTAGE
54% and 70% on three circuits outside the acceptable range can result in
after installing hardware enabling INEFlCIENT OPERATION AND OVERHEATING

LEADING
FDIR capability. TO REDUCED EQUIPMENT LIFE

To date, self-healing technology has been Operating at the upper end of the allowable
installed on only a small percentage of distribu- VOLTAGE RANGE IS A COMMON STRATEGY BECAUSE
tion feeders in the U.S.14 Typically, the least VOLTAGE IS MEASURED AT THE SUBSTATION BUT NOT AT
reliable circuits in distribution systems are first THE END OF THE LINE TO ENSURE THE VOLTAGE AT THE
TO BE UPGRADED BECAUSE IMPROVING THEM WILL BE end of the line is within limits, utilities often set
of the most benefit to customers and the cost of THE SUBSTATION VOLTAGE AT THE UPPER END OF THE
deployment discourages system-wide upgrades. specification. Because load will draw more
Many distribution utilities are now demon- POWER WITH A HIGHER VOLTAGE

TIGHTER CONTROL OF
strating this technology through projects VOLTAGE CAN RESULT IN REDUCED POWER CONSUMP-
PARTIALLY FUNDED BY !MERICAN 2ECOVERY AND tion.iii 4IGHTER VOLTAGE LIMITS AND ITS BENElTS CAN
2EINVESTMENT !CT GRANTSii be effected by a more sophisticated approach to
VOLTAGE REGULATION CALLED hVOLT6!2 CONTROLv
Voltage and Power Flow Optimization WHICH EMPLOYS VOLTAGE SENSORS ON THE LINE

AND
particularly at its end, which feed back the
As customers draw power from the distribution MEASURED VOLTAGES TO THE SUBSTATION #ONTROL
SYSTEM

VOLTAGES DECLINE ALONG DISTRIBUTION EQUIPMENT AT THE SUBSTATION THEN ADJUSTS THE LINES 7ITHOUT INTERVENTION.

THE VOLTAGE LEVEL AT SUBSTATION VOLTAGE AND THE VOLTAGE REGULATING the end of a long distribution line will be lower EQUIPMENT ON THE LINE TO MAINTAIN THE LINE than at the substation. With some exceptions. VOLTAGE AT THE LOW END OF ITS LIMITS THE SIZE OF THE VOLTAGE DROP INCREASES WHEN LOADS INCREASE 4O CONTROL VOLTAGE AT THE &IGURE  ILLUSTRATES THE POTENTIAL EFFECT OF VOLT CUSTOMER LOAD POINT.

UTILITIES HAVE LONG USED 6!2 CONTROL ON THE VOLTAGE ALONG A DISTRIBUTION ADJUSTABLE TRANSFORMERS.

KNOWN AS hVOLTAGE FEEDER )N THE lGURE.

THE VERTICAL AXIS IS VOLTAGE REGULATORS.

v AND hCAPACITOR BANKSvˆA COLLECTION LEVEL AND THE HORIZONTAL AXIS.

FROM LEFT TO RIGHT.

of capacitors that can be switched in and out of represents the distance of a distribution line THE CIRCUITˆTO KEEP VOLTAGES WITHIN A SPECIlED from the substation to the end of the line. range along the entire length of distribution 4HE BLUE LINES SHOW THE VOLTAGE DISTRIBUTION LINES 4HESE DEVICES ARE CONTROLLED FROM THE UNDER NORMAL OPERATION.

which ensures that measurements are not fed back to the substa- THE VOLTAGE FOR EACH CUSTOMER IS BETWEEN  TION 4HE VOLTAGE VARIATION ALONG THE LINE IS SEEN AND  OF RATED VOLTAGE15 &OR RESIDENTIAL TO BE  4HE RED LINES REPRESENT THE VOLTAGE CUSTOMERS. WHERE VOLTAGE distribution substation.

WHERE THE RATED VOLTAGE IS  VOLTS.

DISTRIBUTION USING VOLT6!2 CONTROL 4HE EFFECT THIS GIVES AN ACCEPTABLE RANGE OF – VOLTS of sensing and feedback control is not only to ii 4RACK THESE PROJECTS AT HTTPWWWSMARTGRIDGOV iii 4HIS RELATIONSHIP DEPENDS ON THE NATURE OF THE LOAD 4HE CORRELATION IS VERY STRONG FOR RESISTIVE LOADS.

SUCH AS INCANDESCENT LAMPS.

BUT WEAK FOR MANY ELECTRONIC DEVICES WITH MODERN POWER SUPPLIES THAT COMPENSATE FOR VOLTAGE VARIATIONS Chapter 6: Enhancing the Distribution System 131 .

ANY BEHAVIOR CHANGES18 6OLT6!2 CONTROL ALSO CAN BE APPLIED TO REDUCE FINDING PEAK LOADS BY DECREASING FEEDER VOLTAGE TO THE The use of Volt/VAR control can result lower end of the acceptable range. Figure 6. This in more efficient use of the distribution APPROACH. but IS ESTIMATED TO SAVE THEIR CUSTOMERS BETWEEN ALSO TO TIGHTEN ITS VARIATION ALONG THE LINE TO LESS 1% and 4% of energy consumption without than 4%.1 Voltage Profiles with and without Volt/Volt-Ampere Reactive (VAR) Control 106% Voltage (% rated) 104% 102% Normal Operation Volt/VAR Control 100% 98% Substation ------------.Voltage Regulator 2 --------------> End of Line REDUCE THE AVERAGE VALUE OF THE VOLTAGE WHILE THEIR ADVANCED METERING INFRASTRUCTURE AND maintaining it within its acceptable limits.Voltage Regulator 1 ------------.

KNOWN AS CONSERVATION VOLTAGE network and the possibility of introducing REDUCTION #62 .

CAN YIELD A DECREASE IN THE total power drawn throughout the network. conservation voltage reduction (CVR) thereby deferring the need for capacity expan. sion and increasing efficiency. In a three-year STUDY CONDUCTED FOR  UTILITIES . programs.

ORTHWEST. CUSTOMERS IN THE 0ACIlC .

the practice TAINING THE VOLTAGE BETWEEN  AND  VOLTS AT OF UTILITIES RECOVERING A SUBSTANTIAL PORTION THE CUSTOMER LEVEL EVERY OTHER DAY YIELDED A  of their distribution and transmission costs SAVINGS IN ANNUAL ENERGY DELIVERED BY SUBSTA. As discussed in Chapter 8. MAIN. THROUGH VOLUMETRIC CHARGES.

AS IS CURRENTLY THE tions. Unlike brownouts or rolling blackouts. DOMINANT PRACTICE IN THE 53.

GIVES UTILITIES CONSERVATION VOLTAGE REDUCTION DOES NOT INCENTIVES TO INCREASE THEIR SALES AND DISCOURAGE NECESSARILY NEGATIVELY IMPACT THE QUALITY OF ENERGY CONSERVATION 4HIS INCENTIVE MISALIGN- SERVICE THAT UTILITIES PROVIDE TO THEIR CUSTOMERS MENT COULD SLOW THE INTRODUCTION OF #62 programs that promise reduced energy Many utilities are currently planning to intro. consumption. DUCE VOLT6!2 CONTROL AND #62 PROGRAMS IN THEIR TERRITORY /KLAHOMA 'AS  %LECTRIC 6.2 ADVANCED METERING FOR THE #OMPANY CALCULATED THAT INSTALLING VOLT6!2 DISTRIBUTION SYSTEM CONTROL ON  OF THEIR HIGHEST PRIORITY CIRCUITS WOULD SAVE  GIGAWATT HOURS PER YEAR.

OR .EW CUSTOMER METERING INVESTMENTS ALSO APPROXIMATELY  OF THEIR ANNUAL ENERGY can enhance the performance of distribution SALES.

SYSTEMS 5TILITIES TRADITIONALLY HAVE USED METERS TION ON THEIR EXISTING CAPACITY OF . AND DEFER  MEGAWATTS OF FUTURE GENERA.

strictly for billing purposes. In this traditional watts. MEGA.17 Southern California Edison is deploying PARADIGM.

UTILITY EMPLOYEES VISIT CUSTOMER A #62 PROGRAM THAT USES VOLTAGE DATA FROM premises and manually read electromechanical 132 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .

watt-hour meters that measure electric energy ADVANCED METERING SURVEY REPORTED PENETRATION consumption. These meters are no longer of more than 12 million smart meters by the COMMERCIALLY AVAILABLE IN THE 53.

19 Electronic meters can more tional deployment of AMI systems in a number easily store and communicate energy consump. though many electromechanical meters are 4HE FEDERAL GOVERNMENT HAS SUBSIDIZED ADDI- still in use today. These new meters 2ECOVERY !CT. of utilities through the tion as a function of time. HAVING END OF 24 The difference likely reflects not been replaced by solid-state electronic meters. just growth but also differences in data sources.

metering vary with every utility.25 An estimated of meter reading. to data capture and management: automated MATELY  LARGE DEPLOY- but generally it reduces the costs METER READING !-2 SYSTEMS AND ADVANCED ments throughout the METERING INFRASTRUCTURE !-) SYSTEMS U. These systems typically capture meter readings *UNE  FROM THE STREET USING SPECIALLY EQUIPPED VEHICLES 4HEY CAN REDUCE METER READING COSTS 4HE BENElTS OF ADVANCED METERING VARY WITH by eliminating the need for employees to EVERY UTILITY. The benefits of advanced HAVE ENABLED TWO DISTINCT NEW APPROACHES supporting approxi. improves  MILLION !-) METERS customer support. and enhances AMR technology allows utilities to read customer HAVE BEEN DEPLOYED distribution monitoring and METERS VIA SHORT RANGE RADIO FREQUENCY SIGNALS nationwide as of management.S.

Many utilities throughout OF METER READING. BUT GENERALLY IT REDUCES THE COSTS manually read meters.

IMPROVES CUSTOMER SUPPORT.

THE 53 HAVE DEPLOYED !-2 SYSTEMS OVER and enhances distribution monitoring and recent decades. with more than 47 million MANAGEMENT &OR EXAMPLE.

!-) METERS !-2 METERS REPORTED IN OPERATION AS OF .

!-) SYSTEMSˆALSO KNOWN AS SMART METERSˆ eliminate usage that occurs after customers combine meters with two-way communication HAVE CLOSED THEIR ACCOUNTS. EQUIPPED WITH REMOTE SERVICE SWITCHES ALLOW roughly one-third of the 144 million total U. commercial.21 SERVICE WHEN CUSTOMERS MOVE IN OR OUT 4HIS can reduce the time and cost of connections.S. and industrial meters. utilities to remotely connect or disconnect residential.

Some utility AMI rate filings of recording near-real-time data on power indicate that enhanced detection of theft or consumption and reporting that consumption DIVERSIONˆTHAT IS. AND MAY REDUCE capabilities. These systems typically are capable uncollected bills.

ELECTRICITY USE THAT BYPASSES TO THE UTILITY AT FREQUENCIES OF AN HOUR OR LESS22 A METERˆMAY INCREASE REVENUES /F GREAT Utilities also can typically communicate with IMPORT.

HOWEVER.

IS THE ROLE !-).

AND PARTICU- !-) METERSˆFOR EXAMPLE.

TO CHECK CUSTOMER larly its two-way communication capability. utilities to communicate directly with loads through the meter. can connection status or connect or disconnect play in facilitating dynamic pricing for residen- SERVICE REMOTELY !-) SYSTEMS ALSO CAN ALLOW tial customers. as discussed in Chapter 7. In )N REGULATORY lLINGS. as discussed in Chapter 7.

AMI installations in FOCUSED ON QUANTIFYING THE FOLLOWING OPERA- the U.S. are still in their infancy but are growing tional benefits of AMI: RAPIDLY 0ENETRATION ESTIMATES VARY ACROSS THE INDUSTRY )N . UTILITIES MAINLY HAVE contrast to AMR systems.

DATA SUBMITTED BY THE INDUSTRY s Metering: Installation of AMI means that to the U. This INDICATED THAT  MILLION CUSTOMER METERS IN reduces labor cost and the costs of owning THE 53 WERE !-) DEVICES23 4HE &EDERAL AND MAINTAINING METER READING VEHICLES %NERGY 2EGULATORY #OMMISSIONS  Chapter 6: Enhancing the Distribution System 133 .S. Energy Information Administration meter reading can be fully automated.

s Billing: Billing can be more accurate and and gas businesses that can share much of the TIMELY.

IMPROVING UTILITIES CASH mOWS !-) INFRASTRUCTURE AND OVERHEAD.

increasing accounting ROUGHLY OFFSET THE INVESTMENT COSTS. Where operational benefits on or off remotely. OPERATIONAL SAVINGS MAY COVER A HIGH FRACTION OF !-) s Customer Support: 3ERVICE CAN BE SWITCHED deployment costs.

reducing customer debts. THERE MAY accuracy. and reducing the number of times utility personnel A DECISION TO DEPLOY !-) IS RELATIVELY STRAIGHT- MUST VISIT CUSTOMER LOCATIONS )N ADDITION. and be little or no impact on electricity rates.

At the other extreme. forward. utilities with CALL CENTERS CAN MORE EFFECTIVELY HELP CUSTOMERS RELATIVELY NEW !-2 METERS THROUGHOUT THEIR BECAUSE OF READILY AVAILABLE AND ACCURATE SYSTEM HAVE ALREADY REALIZED MUCH OF THE METERING DATA.

THEREBY REDUCING THE AVERAGE meter-reading cost reductions. the largest single duration of customer support calls. and MAY BE ADDING LITTLE TO THE USEFUL LIVES OF THEIR s Grid Management: 4HE DATA FROM INTERVAL METER SYSTEM THROUGH EARLY REPLACEMENT &OR METERING METERS THAT CAN BE READ AT INTERVALS SYSTEMS IN THIS SITUATION. category of operational benefits of AMI.

precise metering can alert EXTREME CASE ILLUSTRATED IN 4ABLE . In the most for example. OPERATIONAL SAVINGS OF AN HOUR OR LESS CAN BE USED TO MORE may offset half or less of the projected incre- EFFECTIVELY MANAGE DISTRIBUTION SYSTEM ASSETS mental cost of AMI deployment.

OPERA- DISTRIBUTION OPERATORS TO AN OVERLOADED TIONAL SAVINGS COVER ONLY  OF !-) COSTS distribution transformer. )N SOME CASES.

they may not cover the full cost not generate a rate filing and. In rate filings. appear in our table. concerns HAVE BEEN EXPRESSED THAT UTILITIES FEAR OVERESTI- While the operational benefits of installing MATING SAVINGS THAT DO NOT LATER MATERIALIZE OR !-) CAN BE SUBSTANTIAL. THESE BENElTnCOST RATIOS MAY While the operational benefits of installing AMI BE CONSERVATIVE !-) INVESTMENTS THAT CAN BE FUNDED ENTIRELY OUT OF OPERATIONAL SAVINGS MAY can be substantial. do not of the up-front infrastructure investment. therefore.

particularly if the regulator assigns MENT 4ABLE  PRESENTS EXPECTED COSTS AND MOST OF THE RISK TO SHAREHOLDERS. THEY MAY NOT COVER THE INCURRING COSTS ABOVE THOSE APPROVED IN INITIAL FULL COST OF THE UP FRONT INFRASTRUCTURE INVEST. rate filings.

These concerns may bias the country. collected by a search of regulatory REPORTED BENElTnCOST RATIOS BELOW EXPECTED filings. The expected all-in costs of AMI LEVELS DEPLOYMENTS OVER THE PAST lVE YEARS AVERAGED BETWEEN  AND  PER METER ACROSS THESE /PERATIONAL SAVINGS ARE ALSO NOT THE ONLY PROJECTS. AS SOME HAVE benefits for a sample of AMI projects around recently done.

and prior meter AMI-enabled demand response and energy INVESTMENTS 2ECENT DISCUSSIONS OF !-) COSTS CONSERVATION DWARF THE OPERATIONAL SAVINGS28 AND BENElTS FREQUENTLY USE A LOWER RANGE OF 4HEIR IMPORTANCE TO AN !-) INVESTMENT CASE EXPECTED COSTS. labor costs. WITH MUCH OF THE VARIATION DUE TO BENElT OF !-) !DVOCATES OF !-) ARGUE differences across systems in customer mix that the nonoperational benefits from and density.

– PER METER.

THOUGH IT depends on what fraction of costs are offset is not clear how these estimates relate to data BY OPERATIONAL SAVINGS.

reforms. traditional electro. WHETHER UTILITIES AND on actual deployments. and how customers within a particular mechanical meters near the end of their SERVICE AREA ARE LIKELY TO RESPOND 4ABLE  EXPECTED USEFUL LIVES.27 &OR SYSTEMS WITH regulators are willing to commit to pricing dispersed customers.

AND COMBINED ELECTRICITY INCLUDES SEVERAL ESTIMATES OF THESE 134 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .

26 $4–$55 1. 00-E-0165. NY State 2007 1. (TX) “Application of CenterPoint Energy Houston Electric. Connecticut 2010 1.00 California Public Utility Commission.” Maryland Public Service Commission Case No. submitted 07/13/2009.07-07-026.13 $322 $185 0.” Oregon Public Utility Commission.Table 6.” Texas PUC Docket No.19– $63–$804 0. Docket No.20 Delmarva Power & Light Company. submitted 12/14/2010.” Connecticut Department of Public Utility Control. 7/31/2007. 203. Southern 2007 5.1 Operational Benefit–Cost Ratio of Different Advanced Metering Projects Average Operational Nonoperational Operational Total Utility Meters Cost/Meter Benefits/Meter Benefits/Meter Year Benefits/Cost Benefits/ Reference (State) (Millions) Installed Installed Installed Ratio Cost Ratio ($/meter) ($/meter) ($/meter) CenterPoint 2008 2. Baltimore 2009 2.” New York State Public Service Commission Case No. LLC for Approval of Deployment Plan and Request for Surcharge for an Advanced Metering System.58 — — Rochester Gas & Electric Corporation Electric & Gas and New York State Electric & Gas (NY)* Company. Case No. Portland 2007 0. Document No.” Case No. “CL&P AMI Light & Power $484 $232 0. 9207. 9208.58 $159 1. Carpenter and A. UE 189. (MD) “Advanced Metering Infrastructure Business Case and Associated Benefits to Costs Analysis for Maryland in Compliance with Order No. submitted 7/27/2007.75 and Dynamic Pricing Deployment Cost (CT) Benefit Analysis. Portland General 1. 35639.28– B. Docket No.15 — — CenterPoint Energy Houston Electric.09 $253 $128 0. California “Decision Approving Settlement on Edison (CA) Southern California Edison Company Advanced Metering Infrastructure Deployment. 102. 83571. 05-10-03RE01.2 $377– $94– 0. A.33– Connecticut Light & Power.40 Baltimore Gas & Electric Company.50 $478 2.3 $374 $217 0.843 $157 $197 1. Chapter 6: Enhancing the Distribution System 135 . submitted 02/1/2007. Tooman.61 Generation Electric Company.4 $332 $50 0. Delmarva 2010 0. Document No.22 $363 $183 0. “Advanced Metering Infrastructure Overview and Plan.62 2. “Costs Electric (OR) and Benefits.50 $252 1. Gas & Electric “The Smart Grid Initiative Business Case (MD)* Advanced Metering and Smart Energy Pricing Program” Maryland Public Service Commission. submitted 3/31/2010.

” New York State Public Service Commission Case No.90 $36 1. ”Plan for Development and Deployment of Advanced Electric and Gas Metering Infrastructure by Consolidated Edison Company of New York. Inc. .1 Operational Benefit–Cost Ratio of Different Advanced Metering Projects continued Average Operational Nonoperational Operational Total Utility Meters Cost/Meter Benefits/Meter Benefits/Meter Year Benefits/Cost Benefits/ Reference (State) (Millions) Installed Installed Installed Ratio Cost Ratio ($/meter) ($/meter) ($/meter) Rochester 2007 0.3 $243 $218 0. Inc. Consolidated 2007 4. *These projects include upgrades to natural gas metering systems in addition to electricity metering upgrades. 00-E-0165. submitted 3/28/2007. submitted 02/1/2007.Table 6.60 — — Rochester Gas & Electric Corporation Gas & Electric and New York State Electric & Gas (NY)* Company. Inc.” Decision 06-07-027. The number of meters for these projects corresponds to both electricity and gas meters.8 $149 $109 0. Inc.73 $55 1.OTE 4HE DATA IN 4ABLE  HAS BEEN EXTRACTED FROM STATE REGULATORY PROCEEDINGS 0ROSPECTIVE CALCULATIONS OF !-) BENElTS AND COSTS REQUIRE A RANGE OF ASSUMPTIONS EG.05 California Public Utility Commission.10 Consolidated Edison Company of Edison (NY)* New York..67 $250 $150 0.. Electric (CA)* “Final Opinion Authorizing Pacific Gas and Electric Company to Deploy Advanced Metering Infrastructure. and Orange and Rockland Utilities. 2006. 00-E-0165. “Advanced Metering Infrastructure Overview and Plan. July 20.” New York State Public Service Commission Case No. and Orange and Rockland Utilities. Pacific Gas & 2006 9.

PROJECT LIFETIMES.

DEPRECIATION SCHEDULES  !LSO.

AS DESCRIBED IN THE TEXT.

THESE RESULTS ARE SUBJECT TO SIGNIlCANT UNCERTAINTY EG.

TECHNOLOGY COSTS.

DEPLOYMENT SCHEDULES.

DYNAMIC PRICING ADOPTION RATES.

ETC  4HE COST OF !-) PROJECTS ALSO DEPEND ON THE CHARACTERISTICS OF EACH UTILITYS SERVICE TERRITORY AND THE NATURE OF THEIR existing technology and business processes related to customer metering. NONOPERATIONAL BENElTS OF !-) INVESTMENTS )N ADDITION.

these are much more difficult reports called in when electricity is in fact to estimate with precision. Chapter 7 addresses mOWING TO THE METERˆMAY REDUCE COSTS BY these issues in detail. IDENTIFYING hFALSE ALARMSvˆOUTAGE Unfortunately. AVOIDING UNNECESSARY DISPATCH OF CREWS.

although malicious hacking of AMI data could Most utility regulatory filings also omit reli. create such alarms (see Chapter 9 for further ABILITY GAINS FROM BENElTnCOST CALCULATIONS DISCUSSION OF SO CALLED hCYBERATTACKSv  AMI can allow some utilities to respond more rapidly to distribution outages and reduce In the one filing that does incorporate this AVERAGE OUTAGE RESTORATION TIMES FOR CUSTOMERS BENElT.

the utility estimates that the expected MORE RAPIDLY WITH SERVICE VEHICLES. #ONNECTICUT .IGHT  0OWER ESTIMATES !LTHOUGH THE FREQUENCY OR SCALE OF DISTRIBUTION A REDUCTION IN AVERAGE YEARLY OUTAGE DURATION OUTAGES CANNOT BE LOWERED BY USING ADVANCED of six minutes per customer. AMI can enable to “increased notification accuracy that AMI UTILITIES TO QUICKLY AND AUTOMATICALLY PINPOINT PROVIDES DURING STORMSv )N A BASE CASE where an outage has occurred and respond scenario. mainly attributed metering systems alone.

REDUCING THE reduction in outage minutes alone would AVERAGE OUTAGE TIME EXPERIENCED BY CUSTOMERS ACCOUNT FOR  MILLION IN BENElTS OVER THE 136 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .

LIFETIME OF A  MILLION !-) INVESTMENT 6. using reasonable estimates of outage CHALLENGES COSTS FROM A SURVEY OF OUTAGE COST STUDIES29 4HE 2ECOVERY !CT PROVIDED ABOUT  BILLION FINDING FOR THE ELECTRIC GRID.3 PILOT PROGRAMS AND DEPLOYMENT program.

S. $EMONSTRATION 0ROJECTS 3'$0 AND 3MART such as those associated with demand 'RID )NVESTMENT 'RANTS 3')'  4HE GOAL response programs. are a significant factor OF 3'$0 IS hTO DEMONSTRATE NEW AND MORE in determining whether AMI installations COST EFFECTIVE SMART GRID TECHNOLOGIES. nonoperational benefits. Where operational $EPARTMENT OF %NERGY PROGRAMS 3MART 'RID benefits are low. AND MANY UTILITIES HAVE The ratio of operational benefits to total USED THIS FUNDING FOR ADVANCED DISTRIBUTION system technology projects. different utilities. The bulk of this cost for AMI projects varies widely across funding is being distributed through two U.

TOOLS.

TECHNIQUES. are cost-effective.

AND SYSTEM CONlGURATIONS THAT SIGNIlCANTLY IMPROVE ON THE ONES COMMONLY used today.” The SGIG program aims “to &OR SYSTEMS THAT HAVE BEGUN DEPLOYING !-).

Early AND PROMOTE INVESTMENTS IN SMART GRID TECH- AND EFFECTIVE ENGAGEMENT WITH CUSTOMERS NOLOGIES. accelerate the modernization of the nation’s CUSTOMERS CONCERNS ABOUT HEALTH AND PRIVACY electric transmission and distribution systems impacts can pose important challenges.

TOOLS.

”32 Each funding oppor- THAT RADIO WAVES FROM SMART METERS HAVE ADVERSE TUNITY REQUIRED RECIPIENTS TO PROVIDE ONE HEALTH EFFECTS. cybersecurity. AND TECHNIQUES THAT INCREASE appears critical to surmounting these concerns. and While the scientific literature does not suggest operational efficiency. flexibility. situational awareness. interoperability. functionality.

DEBATE OVER THESE RISKS CAN BE SO TO ONE MATCHING FUNDS .INETY NINE 3')' intense that utilities must retool AMI programs RECIPIENTS HAVE BEEN AWARDED A TOTAL OF in response. the  BILLION AND  3'$0 RECIPIENTS HAVE -AINE 0UBLIC 5TILITY #OMMISSION RECENTLY BEEN AWARDED A TOTAL OF ABOUT  MILLION APPROVED A METER SURCHARGE PROGRAM THAT WOULD allow customers to opt out of smart meter Many award recipients are using the funding installation or turn off the wireless transmitting to accelerate distribution modernization FUNCTION. Citing customer concerns.

AND 0ACIlC 'AS  %LECTRIC HAS PROPOSED EFFORTS &OR EXAMPLE.

34!2 HAS BEEN INSTALLING a similar program for its territory. .31 HARDWARE REQUIRED TO SECTIONALIZE CIRCUITS SINCE  PRIOR TO THE 3')' AWARD IN .

.

 0RIVACY ISSUES SURROUNDING !-) ARISE FROM sectionalizing switches already had been the fact that smart meters measure and record INSTALLED 5SING 3')' FUNDING.

34!2 WILL CUSTOMERS POWER USAGE AT FREQUENT INTERVALS UPGRADE FROM AN &$)2 SYSTEM REQUIRING The power usage profiles thus accumulated human action to one that is completely auto- could be used to infer. for example. .ATIONAL SELF HEALING AUTOMATED &$)2 CAPABILITIES Institute of Standards and Technology and BY THE END OF . when a MATED ROUGHLY  OF CIRCUITS WILL HAVE CUSTOMER IS AWAY ON VACATION 4HE .

AND THE UTILITY EXPECTS THAT SEVERAL STATE PUBLIC UTILITY COMMISSIONS ARE  FEWER CUSTOMERS WILL EXPERIENCE OUTAGES ACTIVELY ADDRESSING THESE PRIVACY CONCERNS.

33 3IMILARLY. as a result.

CONTROL OF THEIR DISTRIBUTION SYSTEM SINCE  AND IS NOW USING 3')' FUNDING TO INSTALL $-3 Chapter 6: Enhancing the Distribution System 137 . %LECTRIC 5TILITIES HAS WHICH HAVE GAINED INTERNATIONAL ATTENTION 7E been installing hardware allowing remote discuss the intricacies of this issue in Chapter 9. 00.

distribution utilities with further hardware upgrades enabling full COMMON INTERESTS IN ADVANCED TECHNOLOGIES AUTOMATION OF OTHER 00. appropriately unwilling to publicly share certain and new hardware enabling full automation of information due to concerns about security and ABOUT  CIRCUITS 7ITH MODERN $-3 SOFTWARE THE SENSITIVE NATURE OF SOME BUSINESS INFORMA- and a communications network in place. CIRCUITS ARE EXPECTED may need to use direct channels for coopera- IN FUTURE INVESTMENTS34 tion and communication. a dedicated communication system. software. Indeed. industry STAKEHOLDERS WITH WHOM WE HAVE SPOKEN IN THE 4HE PROJECTS FUNDED BY 3')' AND 3'$0 COURSE OF OUR RESEARCH HAVE INDICATED THAT A MIX PROGRAMS MAY PROVE A BENElCIAL STEP TOWARD of informal and formal mechanisms for DISTRIBUTION SYSTEM MODERNIZATION (OWEVER. tion. Therefore.

THE VALUE OF THESE AND FUTURE ADVANCED DISTRIBU.S. inter-utility collaboration exist in the U. TODAY 3OMETIMES THESE ACTIVITIES ARE COORDI- tion technology deployments and pilot projects NATED VIA SUCH INSTITUTIONS AS THE %LECTRIC 0OWER WILL DEPEND ON EFFECTIVE INFORMATION SHARING.

Research Institute and industry trade associa- THE DEVELOPMENT OF INTEROPERABILITY STANDARDS.

the and efforts to properly align regulatory incen. !MERICAN 0UBLIC 0OWER !SSOCIATION. tions including the Edison Electric Institute.

and lessons Grid Investment Grant and Smart Grid learned is sparse and insufficient to apply to Demonstration Project funding recipients other similar projects. Although industry can SHARE INFORMATION AT SEVERAL NEWLY LAUNCHED and other utilities is essential to capture the WEBSITESˆINCLUDING WWWSGICLEARINGHOUSEORG. AND THE TIVES WITH SYSTEM MODERNIZATION GOALS . expected benefits. Information sharing among Smart utility characteristics. progress. detailed information on federally FINDING funded smart grid projects.ATIONAL 2URAL %LECTRIC #OOPERATIVE !SSOCIATION Currently.

WWWSMARTGRIDGOV. value of these projects.

)NTEROPERABILITY OF )4 EQUIPMENT. AND WWWRECOVERYGOVˆ VERY LITTLE OF THIS INFORMATION PERTAINS TO ongoing distribution modernization efforts.

SOFTWARE.

35 A mechanism for We noted in Chapter 1 a general decrease in sharing the results of publicly funded research. and communications technologies from DESPITE EFFORTS TO INDIVIDUALLY CONTACT FUNDING VARIOUS VENDORS REMAINS AN IMPORTANT CHAL- recipients and collect detailed information on LENGE 4HE . the future grid that emphasizes the importance of information sharing. BUDGETS FOR UTILITY RESEARCH AND DEVELOPMENT DEVELOPMENT. We address these The White House Office of Science and standards and other interoperability concerns 4ECHNOLOGY 0OLICY HAS DRAFTED A FRAMEWORK FOR in Chapter 9. Getting detailed information has been difficult. standards for the future grid that will address these interoperability issues.ATIONAL )NSTITUTE OF 3TANDARDS AND expected costs and benefits for a majority of Technology has undertaken work to coordinate proposed projects.

AND DEMONSTRATION PROJECTS IS ACTIVITIES.

WHICH MAY HAMPER INNOVATIVE ESSENTIAL TO ENSURE THAT THE FUNDING PROVIDED ACTIVITY &URTHER.

FREQUENTLY INVOLVES UNCERTAINTY ABOUT THE Transparency to the public is important not ECONOMICS AND PERFORMANCE OF THOSE INNOVA- only to ensure accountability but also to TIONS #ONVENTIONAL RATE OF RETURN REGULATION facilitate further research that may help guide with ex post PRUDENCY REVIEWS OF CAPITAL FUTURE INVESTMENTS (OWEVER. DEPLOYING NEW TECHNOLOGIES to one utility will benefit other utilities as well.

UTILITIES MAY BE INVESTMENTS MAY DISCOURAGE OR DELAY 138 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .

As these projects progress. Regulators in many states may are an important step toward implementing HAVE CAUSE TO BE SIMILARLY RISK AVERSE TOWARD COST EFFECTIVE ADVANCED DISTRIBUTION TECHNOLO- SIGNIlCANT NEW INVESTMENTS 4HIS PROBLEM IS GIES &EDERAL STIMULUS FUNDING IS CURRENTLY addressed in greater depth in Chapter 8.INVESTMENTS THAT ARE PERCEIVED TO INVOLVE Technology demonstrations and pilot projects greater risks. and OPERATION WILL OCCUR BETWEEN NOW AND  )N THE SITES CURRENTLY HAVE LITTLE USEFUL DATA RELATED large part. are a necessary response to the new challenges these websites may yet facilitate industry THAT ELECTRIC VEHICLES AND DISTRIBUTED GENERATION LEARNING (OWEVER.4 CONCLUSIONS AND to realizing the full potential of these public RECOMMENDATION INVESTMENTS 2EGULATORS ONLY RECENTLY HAVE introduced websites intended to facilitate Substantial changes to distribution system information sharing about these projects. the changes discussed in this chapter to ongoing projects. SUPPORTING MANY SUCH PROJECTS #OMPREHENSIVE inter-utility information sharing will be crucial 6.

It is critical that utilities share TECHNOLOGIES THAT REQUIRE MODERN HARDWARE AND details on both successes and failures. DIRECT UTILITY TO UTILITY pose. software but offer significant benefits in return. Modern distribution management communications and collaboration should systems and other control center software are also be encouraged to ensure that sufficiently necessary to unlock the benefits from increased detailed lessons learned from demonstration deployments of sensors and control hardware. To the extent that these technologies facilitate R E CO M M E N D AT I O N ASPECTS OF PERFORMANCE THAT HAVE NOT BEEN THE Achieving the full potential of federal FOCUS OF TRADITIONAL DISTRIBUTION SYSTEM INVEST- funding for distribution system technology MENTS AND INVOLVE GREATER UNCERTAINTY IN THEIR demonstration projects will require that PAYOFF. PROJECTS ARE AVAILABLE TO INFORM FUTURE INVEST- !UTOMATED &$)2 AND VOLT6!2 CONTROL ARE ment decisions.

smartgrid. Several websites.gov !-) SYSTEMS OFFER CLEAR OPERATIONAL SAVINGS. and failures) are shared widely. UTILITY REGULATORS AND SUPERVISORS MAY NEED NEW APPROACHES TO ENCOURAGE INNOVATION data on those projects (both successes We discuss this at greater length in Chapter 8. including www.

and www.org.sgiclearinghouse. As results become NOLOGY &OR THOSE THAT HAVE NOT. have been AND MANY UTILITIES ALREADY HAVE COMMITTED TO established to disseminate information or executed broad deployment of this tech- about these projects.

energy efficiency programs. Quantifying effectively used to share detailed and the likely magnitude of benefits from AMI. comprehensive data and lessons learned. and increased reliability will be To assure continuation of the modernization critical to assessing the business case for AMI PROCESS INITIATED BY THE 2ECOVERY !CT GRANTS. operational benefits alone may not be sufficient to justify ensure that resources such as these are the cost of a broad AMI rollout. PARTICULARLY available. enabled demand response. policy makers should work to those with existing AMR systems.

INVESTMENTS FOR THESE UTILITIES 7E ADDRESS THESE STATE REGULATORS AND OTHER SUPERVISORS OF IN DETAIL IN #HAPTER  7E DESCRIBE THE PRIVACY DISTRIBUTION UTILITIES MUST BE WILLING TO APPROVE and cybersecurity concerns associated with this INVESTMENTS THAT HAVE A GREATER DEGREE OF RISK technology in Chapter 9. than has been customary for distribution systems. as we discuss at length in Chapter 8. Chapter 6: Enhancing the Distribution System 139 .

REFERENCES # 7 'ELLINGS.

h0OWER $ELIVERY 3YSTEM OF THE 12 &UTURE.

ATIONAL %LECTRICAL -ANUFACTURERS !SSOCIATION AND   n AND % 'ILBERT. no. 12 1 .v IEEE Power Engineering Review 22.

. 'ELBIEN AND " 2OGERS.

!MERICAN .ATIONAL 3TANDARDS )NSTITUTE.

$ /PERATIONAL #HANGE. American h! 4RULY @3ELF (EALING $ISTRIBUTION 'RID 2EQUIRES National Standard for Electric Power Systems and 4ECHNOLOGY !.

v PREPARED Equipment: Voltage Ratings (60 Hertz) 2OSSLYN.

6!.

FOR 'RID )NTEROP  ! 2OAD TO AN )NTEROPERABLE   'RID.

$ENVER.

#/.

.OVEMBER n.

S. Annual Energy - ( 0ERKINS *R.  2 U. Energy Information Agency.

h$IRECT 4ESTIMONY OF -ELVIN ( 13 Outlook 2009 7ASHINGTON.

$# 53 $EPARTMENT 0ERKINS.

*R ON "EHALF OF /KLAHOMA 'AS AND OF %NERGY.

 .

SECTION .

HTTPWWWEIADOE Electric Company.” testimony before the GOVEMEUAERELECTHTML Corporation Commission of the State of /KLAHOMA.

-ARCH .

 3 * ( %TO AND + ( .A#OMMARE.

Electric Power System: 0ERSONAL COMMUNICATION WITH $ -ALKIN. Tracking the Reliability of the U.S.

14 An Assessment of Publicly Available Information * -C$ONALD.

AND " 7OJSZCZYK.

'ENERAL %LECTRIC.

Reported to State Public Utility Commissions 3EPTEMBER .

 "ERKELEY.

AWRENCE "ERKELEY .ABORATORY.ATIONAL . #! .ATIONAL %LECTRICAL -ANUFACTURERS !SSOCIATION AND 15 .

ATIONAL 3TANDARDS )NSTITUTE.   !MERICAN .

SEE NOTE  4 A. Power Distribution System ABOVE Reliability: Practical Methods and Applications + # &AGEN. Chowdhury. A.

h#OST %FFECTIVE %NERGY %FlCIENCY  (OBOKEN.

.* *OHN 7ILEY AND 3ONS.

  -EASURES THROUGH $ISTRIBUTION %FlCIENCY.

EE.v 5 0ERSONAL COMMUNICATION WITH 2OGER .

-ANAGER Metering International    n OF !SSET -ANAGEMENT  3YSTEM 2ELIABILITY.

0ERKINS.

SEE NOTE  ABOVE 17 3OUTHERN #ALIFORNIA %DISON.

3EPTEMBER .

IGHT.  AND &LORIDA 0OWER  .

Florida Power & Light - -ONTOYA.

h3#% )RVINE 3MART 'RID 18 Company’s 2010 Status/Update Report on Storm $EMONSTRATION.

v GRANT APPLICATION SUBMITTED Hardening/Preparedness and Distribution Reliability TO 53 $EPARTMENT OF %NERGY.

.ATIONAL %NERGY *UNO "EACH.

&.

 .

ABORATORY. HTTPWWWPSCSTATEmUS 4ECHNOLOGY .

IN RESPONSE TO FUNDING UTILITIESELECTRICGASDOCS?&0.?REPORTPDF OPPORTUNITY $% &/! .

!UGUST .

  %LECTRIC 0OWER 2ESEARCH )NSTITUTE.

Distribution %LECTRIC 0OWER 2ESEARCH )NSTITUTE.

h4HE !CCURACY OF 19 Reliability Indices Tracking within the United States $IGITAL -ETERS.

v WHITE PAPER 0ALO !LTO.

#!.

 .

0ALO !LTO.

#!.

  HTTPWWWSDGECOMDOCUMENTSSMARTMETER %02)-ETER!CCURACYPDF 7 %TO AND .A#OMMARE.

Energy Information Administration.S. SEE NOTE  ABOVE Ibid. Annual 21 Supply "RUSSELS.  8 Council of European Energy Regulators. 4th Benchmarking Report on Quality of Electricity U.

"ELGIUM.

  Electric Power Industry Report 7ASHINGTON.

$#.

  9 3OUTHERN #ALIFORNIA %DISON.

h 'ENERAL 2ATE #ASE 4RANSMISSION AND $ISTRIBUTION "USINESS 5NIT &EDERAL %NERGY 2EGULATORY #OMMISSION.

2010 22 4$"5 6OLUME ˆ!DVANCED 4ECHNOLOGY.

v Assessment of Demand Response and Advanced PRESENTED BEFORE THE 0UBLIC 5TILITIES #OMMISSION Metering 7ASHINGTON.

$#.

 .

HTTPWWW OF THE 3TATE OF #ALIFORNIA.

.OVEMBER  FERCGOVLEGALSTAFF REPORTS DR REPORTPDF  4 4AYLOR AND ( +AZEMZADEH.

“Annual 23 $-3/-3 )NCREASING $ISTRIBUTION /PERATIONS %LECTRIC 0OWER )NDUSTRY 2EPORT.S. Energy Information Administration. h)NTEGRATED 3#!$! U.

2 7ASHINGTON.v &ORM %)!  Efficiency. no.” Electric Energy T&D Magazine 13.

$# 53 $EPARTMENT OF %NERGY.

  n.

HTTPWWWELECTRICENERGYONLINE   COMPAGESHOW?ARTICLEMAGARTICLE &EDERAL %NERGY 2EGULATORY #OMMISSION.

SEE 24 11 - * 3ULLIVAN.

- -ERCURIO.

AND * 3CHELLENBERG.

NOTE  ABOVE %STIMATED 6ALUE OF 3ERVICE 2ELIABILITY FOR %LECTRIC 53 $EPARTMENT OF %NERGY.

ATIONAL . h2ECOVERY !CT 25 Utility Customers in the United States (Berkeley.ABORATORY.AWRENCE "ERKELEY . 3ELECTIONS FOR 3MART 'RID )NVESTMENT 'RANT #! .

  !WARDS.

140 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .v HTTPENERGYGOVOEDOWNLOADS RECOVERY ACT SELECTIONS SMART GRID INVESTMENT grant-awards-category.

&EDERAL %NERGY 2EGULATORY #OMMISSION.

2011  & # 'RAVES AND ! * "AKER.

Disincentives to Utility  Assessment of Demand Response and Advanced Investment in the Current World of “Competitive Metering Staff Report 7ASHINGTON.

$#.

  Regulation” 7ASHINGTON.

$# %DISON %LECTRIC )NSTITUTE.

 .

HTTPWWWEEIORGWHATWEDO ! &ARUQUI ET AL.

h4HE #OSTS AND "ENElTS OF 3MART 27 0UBLIC0OLICY!DVOCACY3TATE2EGULATION AND -ETERS &OR 2ESIDENTIAL #USTOMERS.

YON.v WHITE PAPER 4 .

h2EGULATORY (INDSIGHT 2EVIEW AND 7ASHINGTON.

$# %DISON &OUNDATION )NSTITUTE )NNOVATION BY %LECTRIC 5TILITIES.

v Journal of FOR %LECTRIC %FlCIENCY.

EEDS. *ULY   AND $ * .

S. Regulatory Economics    n U. Smart Grid Market Forecast.

n '4- 2ESEARCH.

  ! &ARUQUI AND . 7OOD.

” presentation at the . h)%% 2ELEASES 4HE 28 Benefits of Smart Meters.ATIONAL !SSOCIATION OF 2EGULATORY 5TILITY #OMMISSIONS 7INTER -EETINGS.

7ASHINGTON.

$#.

&EBRUARY n.

IGHT  0OWER.  #ONNECTICUT .

h#.0 !-) AND 29 $YNAMIC 0RICING $EPLOYMENT #OST "ENElT !NALYSIS.

O   2%.v $OCKET .

O . #OMPLIANCE /RDER .

-ARCH .

.

HTTPWWWDPUC STATECTUSDOCKHISTNSF7EB -AIN 6IEW3EARCH %LECTRIC/PEN6IEW3TART+EY  2% California Council on Science and Technology.  Health Impacts of Radio Frequency Exposure from Smart Meters 3ACRAMENTO.

#!.

  AND -AINE #ENTER FOR $ISEASE #ONTROL.

h-AINE #$# %XECUTIVE 3UMMARY OF 2EVIEW OF (EALTH )SSUES 2ELATED TO 3MART -ETERSv !UGUSTA.

-%.

  -AINE 0UBLIC 5TILITY #OMMISSION.

h-05# $ECIDES 31 3MART -ETER )NVESTIGATION.

v PRESS RELEASE.

-AY .

 AND - #HEDIAK.

State Regulator Says. -ARCH .” Bloomberg. h0'% 3HOULD !LLOW 3MART Meter Opt-Out.

.

53 $EPARTMENT OF %NERGY. HTTPWWWBLOOMBERGCOM NEWS  PG E SHOULD ALLOW SMART METER opt-out-state-regulator-says-1-.html.

h/VERVIEW OF 32 0ROGRAMS.

3TUDIES AND !CTIVITIES.

v HTTPWWW SMARTGRIDGOVRECOVERY?ACTOVERVIEW 'ILBERT.

'ELBIEN.

AND 2OGERS.

34!2. SEE NOTE  ABOVE 33 AND .

34!2 %LECTRIC %XPANDS 3MART 'RID 4ECHNOLOGY 4HROUGHOUT 3ERVICE 4ERRITORY TO )MPROVE 0OWER 2ELIABILITY FOR !LL #USTOMERS. h.

v PRESS RELEASE.

!UGUST .

 0ERSONAL COMMUNICATION WITH 4IM &IGURA.

0ROJECT 34 -ANAGERˆ3MART 'RID.

00.

&EBRUARY .

 /FlCE OF 3CIENCE AND 4ECHNOLOGY 0OLICY.

A Policy 35 Framework for the 21st Century Grid: Enabling Our Secure Energy Future 7ASHINGTON.

$# %XECUTIVE /FlCE OF THE 0RESIDENT.

  Chapter 6: Enhancing the Distribution System 141 .

142 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .

Section 7. Complementary technologies that promise to automate customer responses are likely to be important in these objectives. Section 7. investment costs. Carefully designed customer engagement programs and transition policies to dynamic pricing tariffs will be critical to creating responsive demand and realizing the full potential of advanced metering infrastructure (AMI) investments. and energy. We first recommend that utilities that have already committed to AMI deployment prioritize transition paths to broad-based dynamic pricing. and shifting Policy makers.3 then assesses the potential benefits of more active demand management by electricity customers. Technologies that facilitate increased demand response and enhance energy efficiency are likely to become more affordable as they mature and costs continue to fall.” INTRODUCTION reduce the need for expensive investment in additional generating capacity. More responsive demand can improve reduce the total cost of delivering a given system efficiency and reduce costs. we discuss the opportunities for grid operation associated with more actively engaging electricity demand. while Section 7. particularly at the residential level. consumption to off-peak periods can flatten utilities. Such data can improve program design for adopters and decision quality for utilities that must evaluate whether to accelerate replacement of their customers’ meters.Chapter 7: Engaging Electricity Demand In this chapter. particu.4 describes energy conservation and energy-efficiency programs and their relationship to demand response programs. Real-time adjustments to consumption at or near system peaks can demand could reduce the cost of managing Chapter 7: Engaging Electricity Demand 143 . larly to variations in the cost of supplying the system.1 introduces the motivation for increasing demand engagement. we note that results of ongoing pilots and system deployments could dramatically narrow uncertainty surrounding the costs and benefits of AMI. The past several years have seen significant and growing interest in encouraging greater customer involvement in managing their electricity use. Section 7.5 provides our main findings related to demand engagement. and results of demand response initiatives. Reduced quantity of energy. We recommend that decision makers for utilities confronting uncertain investment cases recognize the option value of deferring decisions pending that early deployment data: adoption decisions are not “now or never” but “now or re-evaluate with new data. Sections 7.6 details our recommendations. utilization on system efficiency and reduce costs. increasing interest in electricity load becoming improve capacity More responsive demand can improve more responsive to system conditions. Finally. Their relative immaturity in 2011 signals the importance of ensuring upward compatibility and interoperability.1 through 7.4 provide background. and customer groups have expressed load curves. Section 7. Section 7. and assesses the substantial uncertainty that continues to surround estimates of potential demand response benefits.2 describes demand response programs as they exist today. grid system operators. to avoid stranding customer investments. We highlight the importance of publicly sharing information on their customer engagement programs. regulators. and dynamic pricing—if data are shared. associated technologies.

vehicles and plug-in hybrid vehicles. has seen have reached the end of their useful lives by dramatic growth in the deployment of 2030. making their own real-time electricity usage more visible to consumers may complement AMI permits fine-grained communication energy conservation goals. the economic case for AMI may. involve even modernization and automation program.S. For potential early “interruptible” power in exchange for lower adopters. Some operate in providing more granular data on customer states that have made universal AMI deploy- usage. of system conditions to customers and fine- grained measurement of customer responses. these operational benefits of AMI consumption and facilitate their responses to frequently constitute only a fraction—for some price or other supply-side signals. particularly those that have installed automated meter reading systems. by promoting off-peak charging. and some form of AMI is almost certain advanced metering infrastructure (AMI). To capture these potential conditions—holds particular promise. thermostats and “smart charging” of electric as part of a broader distribution system vehicles—that can. For others. AMI complements a range of “smart” ment a key policy objective. however. U.S. supply volatility brought on by increased must convey more than simple signals of penetration of variable energy resources emergency conditions. benefits. could reduce the need for new generation and network capacity to accommodate electric As we look to 2030 and beyond. even before their current tion with the distribution utility. In addition to meters have fully depreciated. More it is difficult to envision a scenario price-responsive demand may help to mitigate without broad deployment of some market power concerns in restructured whole- sale generation markets by reducing the form of advanced metering across profitability of price increases. as energy response and management technolo. therefore. The growing interest in engaging electricity As we look to 2030 and beyond. In addition. technologies capable of facilitating this change. AMI deployment. New technologies make participation in electricity markets. only the demands of those small fraction—of the estimated capital costs large customers who had agreed to take of its universal deployment. Most currently installed meters will As described in Chapter 6. respond appropriately to those communications. system-wide deploy- gies—such as programmable controllable ment of AMI may reduce operating costs and. But records a customer’s electricity consumption at utility systems may choose to invest in “early” least hourly and provides two-way communica. and customers must (VERs) and. which to be the replacement technology of choice. it is difficult to demand has been accompanied by significant envision a scenario without broad deployment advances in the development and diffusion of of some form of advanced metering across U. However. communication to customers 144 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .S. rest heavily on the magnitude of and then only in emergencies or when demand benefits it creates through greater demand-side nears system capacity. described in Chapter 6. utilities. as in response to contemporaneous system suggested above. as noted in in more active management of their electricity Chapter 6. in principle. the U. utilities. utilities. smaller commercial and residential customers enhance service quality. rates were responsive to system conditions. AMI’s it easier to reduce the demands of even smaller ability to facilitate demand response through customers in emergency conditions. But the dynamic pricing—tariffs that go beyond simple potential benefits of demand response go time-of-use prices by allowing prices to change well beyond dealing with emergencies. a relatively Historically.

and supply and demand at each moment in time.3 It also reflects concerns about customer reactions As we have discussed in earlier chapters. particularly have not yet committed to near-term AMI deployment can use this information to decide Electricity on average accounts for only when and how to make cost-effective use of 2% to 3% of household expenditures.4 Chapter 7: Engaging Electricity Demand 145 . the differential impact of dynamic pricing on System operators historically have ensured the low-income customers. An with their greater metering cost. inform smaller commercial and residential customers “all-in” cost estimates for system-wide adoption. Pennsylvania. This chapter discusses equality of supply and demand largely through the nature and impact of current demand supply-side tools: operators adjust the dispatch response programs and assesses the evidence schedule for generators to meet forecast on the potential for expanded demand engage. including California. residential and smaller commercial customers. The discussion highlights an important ques. day. Predictable variation in demand over time. demand. can help refine Yet regulatory commitments to tap the capa. adjust dispatch as forecasts are ment. updated. if broadly shared. responsive the typical smaller consumer is likely Utilities and public utility commissions that to be to signals of electricity costs.1 The focus estimated 27 million AMI meters have been of recent research therefore has been on deployed nationwide as of September 2011. the wide range of estimated potential benefits bilities of these meters for dynamic pricing by of AMI-facilitated demand engagement.2 these projects. and Texas.1 WHY ENGAGE DEMAND? 2% to 3% of household expenditures. i The 2015 projections may be aggressive given that AMI deployments frequently have been delayed or deferred. often lag their installation. and some projections suggest installations Pilot projects demonstrate that experimental could reach more than 60 million (40% of populations of residential customers reduce utility customers) by 2015. assisted by legislative directives in some commercial customers.i Information from consumption when electricity prices are high. particularly by residential customers. to AMI technology. advanced meters. tion for regulators and utilities: what are the and season is associated with changes in the trade-offs between rapid universal rollout of marginal cost of supplying electricity that AMI technology and a staged deployment that regularly vary by a factor of two or three. and smart grid stimulus grants under the The cost-effectiveness of dynamic pricing American Recovery and Reinvestment Act of already has been demonstrated for large-scale 2009 that have committed utilities to system- industrial and commercial customers. volatile electric bills electricity systems must precisely balance (particularly unexpectedly high bills). states. and use ancillary services to adjust to real-time deviations in demand from forecast.The cost-effectiveness of dynamic evaluates the experience of early adopters pricing already has been demonstrated before embarking on later waves of installations? The diffusion of AMI meters has been growing for large-scale industrial and rapidly. This appears to be and highlight both technical and human-factor due in part to continuing debate over how challenges in rolling out these technologies. even wide installation of AMI infrastructure. since electricity on average accounts for only 7.

as Chapter 1 has shown. time. unanticipated changes in prices that may not vary even seasonally. the PJM Interconnection region 2010 “peakedness” of demand—measured.com/markets-and-operations/ energy/real-time/monthlylmp. or result.” http://pjm. the majority tion networks because all must be sized to meet of electricity customers in the U. and consumption can be highly August afternoon—may cause the marginal “peaked. “Monthly Locational Marginal Pricing. including predicted maximum demand (plus a safety almost all smaller commercial and residential margin) at all times and across all geographical customers.aspx.5 And. Compare.S.1 with the hourly prices (inset) for over a year—generally has been increasing over two summer 2010 days.. Some face Figure 7. This problem affects generating capacity as well as electricity transmission and distribu- Despite these cost fluctuations. As a generator or transmission availability.760 hours) account (MWh) of electricity to consumers to increase for 10%–18% of the capacity needs in North fivefold or tenfold or more.” It is estimated that fewer than 1% of cost of providing an additional megawatt hour annual hours (60–100 of 8.1 2010 Average Daily versus Selected Hourly Average Real-Time Locational Marginal Price (RT LMP) in the PJM Interconnection Hourly RT LMP – Jul 7 and Aug 11 400 200 350 300 250 180 200 150 160 100 50 0 140 1 3 5 7 9 11 13 15 17 19 21 23 24 Hour LMP ($ / MWh) 120 100 80 60 40 20 0 1–Jan 15–Jan 29–Jan 12–Feb 26–Feb 12–Mar 26–Mar 9–Apr 23–Apr 7–May 21–May 4–Jun 18–Jun 2–Jul 16–Jul 30–Jul 13–Aug 27–Aug 10–Sep 24–Sep 8–Oct 22–Oct 5–Nov 19–Nov 3–Dec 17–Dec 31–Dec Date Data Source: PJM. the instance. face retail prices that do not change locations. More important. air shift their electricity use away from high system conditioning loads on an exceptionally hot demand hours. say. over the course of a day or a week. 50 50 50 50 146 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . these customers have no incentive to unusually high demand—for example. by average daily real-time wholesale prices shown the ratio of peak demand to average demand in Figure 7. for America.

however.Demand response programs may reduce the and encourage consumers to respond (either total cost of maintaining system balance by manually or automatically) to price or other inducing changes in consumption. while Demand responsiveness is jeopardized. and there is considerable uncertainty in estimates of the likely net In practice. and thus the opportunity cost of energy. “price-responsive demand” results as clothes drying.” 6 efficiency and conserva. other. though they may be related. customer load response may differ impact. Whether the direct electricity at less than the observed current. granular consumption data versus reactive programs. and though from their normal consumption patterns in they may be related. weekday afternoon. or to install a effect of turning up an air conditioner thermo- switch that gives the operator or a third-party stat or turning off lights during a summer aggregator direct control of a customer’s load. this definition of demand response: “Changes Demand responsiveness is distinct from energy in electric use by demand-side resources efficiency or conservation. goals. The Federal A second reason often given to engage energy Energy Regulatory Commission (FERC) uses demand is to reduce energy consumption. “Load management or Some conservation may one does not imply the other. For example.8 But some peak use. utilities. or third-party provide additional support to conservation aggregators directly control load responses. such Alternatively. It may be useful going forward to The detail and immediacy of energy consump- distinguish between dispatchable programs. is high. different mechanisms. building energy use and reduce consumption. when customers face prices that vary with and lower off-peak prices associated with system supply conditions and decide to reduce many dynamic pricing structures may further or to shift consumption when they value increase off-peak usage. or to do both). may which system operators. based on customers’ may offer insight into opportunities to tune voluntary reactions to price or other signals. and this market the attractiveness of dispatchable load programs seems likely to see substantial growth Chapter 7: Engaging Electricity Demand 147 . empirical question.distinct from energy efficiency tion programs target or conservation. and As the FERC definition suggests. may simply be rescheduled. as when usage at peak periods is expected level in response to an instruction or eliminated rather than shifted: consider the signal from the system operator. control” programs offer customers incentives to occur through demand reduce their consumption below a baseline or response. these reductions in overall programs generally rely on one of two very electricity consumption. in tion data from AMI meters. net effect of demand response is to reduce or period price (or to increase consumption increase overall consumption is ultimately an during low-price periods. when system capacity utilization. These need not be mutually exclusive at the A variety of third-party providers currently offer level of the utility or competitive retail supplier services to optimize building energy use based of energy: dynamic pricing may both enhance on detailed energy usage data. one does not imply the response to changes in the price of electricity.7 Demand response programs generally or to incentive payments designed to induce focus on reducing consumption at particular lower electricity use at times of high wholesale times. often shifting it market prices or when system reliability is to other times. substantially within and across these mecha- nisms. however. particularly signals they receive. however.

this may involve industrial customers. demand response programs have grown tion. for example. These mechanisms provide information substantially in recent years.9 In the since the 2006 survey.3.g. More recently. electricity consumption. who account for roughly either direct real-time information on consump. email.S. (e.” 12 148 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .2 Informational Feedback Continuum 1 2 3 4 5 6 Standard Billing Enhanced Billing Estimated Daily/Weekly Real-time Real-time Plus (e. research increasingly focuses on the TODAY use of behavioral interventions through feed- back mechanisms to reduce electricity consump. surveys (released in 2007. 60% of U.. 2008.S. for example. appliance by mail. household (e. analyze the enrolled across a wide range of demand impact of reporting a consumer’s electricity response programs. often in the context survey date).2 DEMAND RESPONSE PROGRAMS applications. Figure 7. as shown in Figure 7.S. as AMI deployment expands. 2009.ii In residential 7. programs that allow end of the continuum.g.2. recorded 53 gigawatts of potential peak load vation. demand response programs have grown collected in its 2006.11 Commercial and residential electricity sector. in-home networks. By the 2010 survey. such as energy conser. signal capability) and/or control) est. but could also contribute customers or third-party aggregators to bid more granular information to indirect feedback demand response into wholesale markets have mechanisms. web-based (e. home-area bi-monthly) advice. and 2011. monthly. FERC reports enrollment in demand response programs U. appliance specific energy audits and consumption displays. U. based on (e. FERC of positive social norms. 2009). The utility promises participating sites “10% or more savings on their total addressable energy expenditures. and reporting data for the full year prior to the and assessment of behavior.10 increased dramatically and now comprise the largest category of demand response by enrolled load.. They might. Residential Electricity Use Feedback: A Research Synthesis and Economic Framework (Palo Alto.g. disaggregation) self-meter) Indirect Feedback Direct Feedback (provided after consumption occurs) (provided in real-time) Information Availability Low High Cost to Implement Source: Electric Power Research Institute. forming a continuum. Figure 7. info and Feedback Feedback Feedback (e. pricing. and 2010 biennial substantially in recent years. AMI would be an integral part of the primarily through various load management systems of direct real-time feedback at the upper programs. CA...g..g. measurements. to support AMI-enabled commercial and institutional customer access to “EfficiencySMART Commissioning” software.g.. reflecting an 80% increase usage in comparison to that of neighbors. disaggregation or otherwise) billing analysis. ii Southern California Edison recently entered a contract with the energy management company EnerNOC. have tion or indirect feedback provided at some later historically accounted for the bulk of load time. as illustrated in enrolled in demand response programs.

0 mandatory response programs.3 Reported U. reactive (defined 4.S.000 10. The North American extended to nondispatchable. customers may not be drawing their full enrolled load at the time of a demand response event.5 Electric Reliability Council (NERC) is to include price-mediated) demand response developing 4.13 At present.0 responses to system “events” in use of demand response resources to meet one an effort to improve data accuracy and or more core objectives. voluntary demand response programs typically calls on a given provide less predictable and lower overall responses per MW of enrolled load than do direct load control or 0.5 response enrollment and collect information industry confidence in the measurement and on their actual 3.iii There is some promise of events.14 FERC reports that aggregate actual demand response averages less than one-third of potential peak reduction. 50 50 50 50 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Measuring demand response potential is predictability of Measuring demand response challenging.5 to actual.0congestion may render demand response from customers in uncongested areas irrelevant. 2006 Assessment of Demand Response and Advanced Metering Staff Report (Washington. 2008). FERC’s “potential peak load demand response potential is challenging. Potential Peak Load Reduction by Customer Class in 2006. 2010 Assessment of Demand Response and Advanced Metering Staff Report (Washington. and 2010 Federal Energy Regulatory Commission (FERC) Surveys (megawatts) 25. DC. 2011). Nevertheless. this will be progress on this front. 2006).15 Chapter 7: Engaging Electricity Demand 149 . 2008.000 15. or available load reduction at any given time for a number of reasons. For example. DC. Federal Energy Regulatory Commission. 1926 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 reduction” is a better indicator of aggregate resources.000 5. Finally. Figure 7. at peak periods.5 customer to reduce their load. In Phases III and IV.0 iii Enrolled “potential peak load reduction” as measured by the FERC survey does not necessarily translate 1. Federal Energy Regulatory Commission. DC. 2. it currently controllable resources that system operators provides the greatest comparability over time can direct to respond to reliability-driven and across systems. local transmission1.000 0 Commercial Residential Wholesale Other & Industrial 2006 FERC Survey 2008 FERC Survey 2010 FERC Survey Source: Federal Energy Regulatory Commission. NERC’s goal is to develop Data System to codify reporting of demand performance-based data that will enhance 3.000 Potential Peak Load Reduction (MW) 20. with substantial variation in that figure across regions and years. load enrolled in a demand response program NERC is focused in than of expected reductions in system demand Phases I and II only on dispatchable. 2008 Assessment of Demand Response and Advanced Metering Staff Report (Washington. expected.0 a Demand Response Availability resources. or restrictions in the program may explicitly or implicitly limit the frequency or duration of system operator 0.5 2.

programs. as shown in Figure 7. The immediate and often predict- The greatest growth in programs targeted at able load reductions many of these programs peak load management has been in regions that provide are especially valuable in responding to allow demand-side resources to participate in system emergencies. In More recent programs have begun to focus on reducing addition. vertically efficient investment in and use of capacity. customers typically receive a payment or rate tion services or storage.16 More recent programs have begun to focus on The designs of demand response programs reducing consumption in non-crisis peak used to accomplish these objectives are varied. such as with this aim elicit customer reductions. transmis. and system operators to these may complement reliability objectives. and regulations that allow for an the late 1960s by load-serving entities. generation. The value of this respon. periods and smoothing demand over time. quick common focus has been reliability. particularly for purely price-mediated demand response. in response cles. in the capital-intensive control. and distribution capacity that is sized to meet the maximum expected demand may go Load Control Programs underutilized or unused in most periods. Demand response programs may be targeted VERs and the importance of loads that may be at various objectives. such as wind or solar genera. While integrated utilities. at scale for this purpose may vary widely across tricity demand threatens to exceed the available types of programs.17 especially for regions with significant variable While programs differ in operational details energy resources. credit as an incentive for participation. by shaving the increasingly sharp peak loads described in Chapter 1 that are associated with Finally. raising the long-run average cost of supplying a given Load management or control programs amount of energy.4. The feasibility of using demand response to system emergencies or periods when elec. supply of generation or network capacity. air conditioning and charging of electric vehi- usually through load management. Solutions to this “peak load” comprise the largest segment of demand problem focus on economic incentives. if loads could be three categories: emergency payment. As They can be broadly categorized into load noted in Chapter 1. changes to market rules and reliability standards are required in some regions to allow consumption in non-crisis peak periods and smoothing demand-side resources to participate fully in the demand over time. sion. inter- quickly and reliably adjusted. objectives. provision of balancing services. there may be opportunities going increased penetration of air conditioning. and customer price–mediated limited storage capability. they generally fall into tors. They can also reduce costs long-term wholesale capacity markets. Enrolled might substitute for flexible supply-side regula. from entity to entity. wholesale market–administered electricity industry with no inventory and programs. The earliest and most especially amenable to predictable. maintain reliability. Used since policies. Load siveness is likely to increase with penetration of management programs contributed 150 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . remains to be demonstrated. demand response ruptible tariffs. and direct load control. and operational feasibility. price response programs in place today. they have more recently their focus tends to be on long-run efficiencies begun to incorporate broader peak-shaving rather than short-run operational exigencies. forward to use demand-side resources to help falling industrial load shares. Programs responses through automated controls. As noted in Chapter 3. that do not vary with system conditions. and retail prices balance energy supply and demand in real time.

000 6.000 4. approximately 62% of the nation’s potential incentive payments—ranging from $150/MWh peak load reduction in FERC’s most recent to $1. Assessment of Demand Response and Advanced Metering Staff Report (Washington. markets.000 Enrolled Load (MW) Commercial & Industrial Residential Wholesale Other Retail Source: Federal Energy Regulatory Commission.000/MWh of curtailed energy use—to survey and continue to grow. 2011).18 The aggregate customers who curtail demand during reliability load (MW) enrolled in different categories events. and $1. provide the grid operator or load-serving entity with the same degree of certainty of load “Emergency programs” are now the largest reduction as direct load control or interruptible category of demand response program by tariff programs.000 14. propelled to this position for 25% of national peak reduction potential by a dramatic increase in wholesale demand reported in 2010. curtailment in emergency of load control programs is summarized in programs is generally voluntary and does not Figure 7. Midwest.102/MWh in the Southwest Power Pool.4 Enrolled Load by Type of Demand Response Program and Customer Class Interruptible Load Direct Load Control Load Controlled Emergency Demand Response Demand Bidding and Buy-back Load as a Capacity Resource Spinning Reserves Non-spinning Reserves Market Administered Regulation Other Critical Peak Pricing Critical Peak Pricing with Load Control Peak-time Rebate Price Mediated Real-time Pricing System Peak Response Transmission Tariff Time-of-use 0 2.4. Emergency programs accounted enrolled capacity. These programs provide iv These are in the range of maximum wholesale prices for generation in most U.000 10.Figure 7. and Electric Reliability Council of Texas regions. For example.000 8.000 12.S.iv However. Chapter 7: Engaging Electricity Demand 151 .20 Wholesale prices provide an appropriate benchmark for incentive payments because they reflect the cost of energy inclusive of grid conditions. 2010 maximum wholesale prices were $396/MWh. $343/MWh. DC.19 response enrollment. respectively.

peak shaving on a small number operator through switches or meter controls. participation has been of the country in that year.22 With loads under the events during the summer of 2001 range from direction of the system operator. In some cases. 14 out of 30 potential peak load reduction.” These programs focus primarily on provide additional targets for enhancing large commercial and industrial customers. benefits over the entire summer totaled over ability of response. But when needed. For example. and respondents indicated plans to substan- tially increase load enrolled in these programs.4 million. third-party aggregators Most load management programs are may be able to combine smaller users to provide interruptible demand. direct capacity involves “interruptible or curtailable load control programs for charging stations may tariffs. The amount of constitute the largest category of demand curtailment needed for reliability appears to be response program by number of customers consistently much lower than available peak enrolled.21 Most load management programs are used relatively infrequently because they are designed “Direct load control” programs. New York of residential peak load nationwide. and costs vary widely predicted to increase by 6. are the two Independent System Operator’s estimated most common targets of residential direct load reliability benefits for each of four curtailment control programs. Respondents to FERC’s $20 million. number of hours during the year. system conditions. For instance.25 It is difficult. programs achieve predictable load reductions particularly for direct load control when called.300 MW by 2015. The second-largest category by enrolled As electric vehicle penetration increases. 152 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . demand response potential.24 In the 2010 FERC particularly strong among residential customers. even though record- 2010. Direct load load management programs surveyed in 2001 control accounts for approximately 17% of the were reported to have operated just once or not national peak reduction potential reported in at all during that year. to measure 2010 survey report the greatest expected growth the dollar value of reliability-related peak in direct load control programs over the near reduction on a nationwide basis. in which to provide reliability under emergency condi- customer equipment or appliances are directly tions or. reducing consumption during declared emer- gencies. and the third largest by total enrolled reduction potential. with enrolled load in these programs its use. which comprise 70% considerable value. Although these programs may be offered setting peaks were experienced in some parts to all customer classes.23 across the country. Central air conditioning cycling and tion. survey. Customers may face used relatively infrequently because significant penalties if they fail to comply with they are designed to provide reliability a curtailment signal. see the discussion who receive a rate discount or credit for in Chapter 5. however. which ensures that these under emergency conditions or. particularly for direct load control controlled by the load-serving entity or grid programs. these programs create electric hot water heaters.000–$3. of hours during the year. particularly as term. and aggregate reliability provide dispatchable load with high predict. respondents report actual reductions with more than 5 million enrolled residential equivalent to 30% of the potential peak reduc- customers. peak shaving on a small approximately 20% of national peak reduction potential reported in the 2010 FERC survey. these programs $800. Interruptible tariffs comprised programs.

recommends paying the wholesale market price ahead or hour-ahead programs for near-term for reductions from baseline. sale programs on behalf of their customers.26 customers can subscribe to services offered by Programs administered by independent system third-party aggregators or “curtailment service operators (ISO) or regional transmission providers. 745 “energy price programs” typically are day. customers are not committed to purchase their In many wholesale markets. particu- programs explicitly or implicitly designed to larly in the Electric Reliability Council of Texas call for load reductions over a low number of region. non. “Capacity Comverge aggregated almost 8. enables them to realize net savings even after bearing the costs of advanced meters and Wholesale Market-administered Demand control equipment.30 expanded as large customers bid into forward capacity markets. as illus. These factors for demand response has emerged over the past enhance their attractiveness for increasing decade and is gathering momentum. They can be categorized demand reduction should the utility or system as capacity. participate in whole- organizations (RTO) contributed approxi. The aggregator is responsible for finding the tial reported in 2010. Smaller trated in the bottom panel of Figure 7. Ancillary service programs limits the frequency of their use. demand responsiveness in aggregate. in turn. committing to future load Uneconomic behavior can be encouraged when reductions when called as a substitute for some of these programs are layered on top of (unbuilt) new generating capacity.32 Customers facing high elec- spinning reserves. in 2010. which tricity prices generally will react by reducing are used to ensure the balance of supply usage. the load enrolled in wholesale programs. FERC Order No.29 mately 27% of national peak reduction poten.4. In contrast. and regulation. This may be a and participate directly in these programs. For example. programs represent a very small but growing most operational experience has been with share of wholesale demand response. paying them in addition for those Chapter 7: Engaging Electricity Demand 153 .000 MW of market programs” are designed to incent residential load through a variety of programs market participants to commit to load reduc. dynamic pricing tariffs.” on an energy price basis. energy price or demand bidding.” who.While nothing intrinsic in incentive-based and demand in real time. customers are baseline quantities of electricity. either through long-term capacity tions as a substitute for increased generating contracts or programs it managed on behalf capacity. This makes Commercial and industrial customers along it difficult to extrapolate participation and with wholesale purchasers comprise most of response rates should the programs be broad. One motivation for voluntary load reductions during emergencies this has been concern about “demand fatigue. operator call for load to be shed. for reductions. doing so increasingly allowed to provide “ancillary generally will provide excessive compensation services. fruitful area to explore further in well-designed Their volume of electricity consumption pilot experiments. and they frequently are Response Programs better able to commit to transactions required for participation in wholesale programs.28 decreased customer responsiveness or exit from the program if called too frequently. and ancillary service programs.” such as spinning reserves.31 but when commitment to consumption reductions. Large ened to pursue objectives other than reliability. either A variety of new wholesale market programs directly or through aggregators. commercial and industrial customers can enroll such as peak load management.27 Some ISOs and RTOs also compensate days or hours of the year. Enrollment in these programs has of utilities or other clients.

They do not. they will both reduce their peak pricing. while the price schedule reflects the fact that costs are on average higher during periods in which demand is on average higher. This predictable and stable time profile of prices can encourage customers for reducing consumption from customers to make long-term changes in their a baseline generally provide excessive consumption patterns—for example. Aggregate enrolled load If customers have not committed in advance across all customer price-mediated programs to purchasing a quantity of electricity or a contributed only about 8% of the national particular load profile. FINDING and a higher peak rate for use during weekday Demand response programs that pay afternoons and early evenings. principle underlying these programs is that customers will adjust their usage to consume Dynamic prices can respond to those condi- electricity only when the value they ascribe to tions. and weekend hours. and peak-time rebates. The generating unit.1 million residential baseline or act on their private information to customers and 250. early mornings. determining the amount enrolled peak load potential reported in FERC’s of demand response eligible for payment also 2010 survey.4. usually held constant over a season. industrial customers. response. Price-Mediated Demand Response Programs time-of-use prices are not truly dynamic. reductions amounts to double compensation. accounting for approxi- mined baseline and their true consumption mately 4% of the enrolled potential peak load patterns. incentives for strategic behavior. common time-varying pricing structures are They will. demand reductions that may be promised For example. programs are much smaller than those in the previous categories. These electric bills and be paid for doing so. The four most the spectrum is “real-time pricing. for example. by compensation and give customers avoiding dishwasher use during early evening. However. the greatest promise for future demand tion in the program is inherently unknowable. but dynamic pricing plans vary widely in its consumption is higher than their cost of the frequency of price changes.” incentives for customers to manipulate the These enrolled roughly 1. What residential sector. But these programs may hold would have been purchased but for participa. be selling something they time-of-use prices. Wholesale programs that compensate demand for reductions in load relative to a historic or As noted in Figure 7. At one end of purchasing that electricity. critical have not bought. real-time pricing. set a low off-peak rate that applies to weekday nights. distinguish normal July Price-mediated demand response programs weekday afternoon conditions from those on face customers with retail prices that change an unusually hot day with a spike in air condi- over time to reflect variations in the market or tioning loads or an unexpected outage at a large opportunity cost of providing electricity. the most common form administratively determined baseline provide of time-varying prices is “time-of-use tariffs.” in which 154 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .33 Demand participation in long-term across all demand response programs. in effect.34 Time- capacity markets also raises questions of the of-use rates set a time profile of prices far in verifiability and credibility of contracted advance. with very little of that in the requires a difficult counterfactual analysis. summer time-of-use rates might several years in the future.000 commercial and profit from differences between the predeter.

But. except that customers are particularly of real-time pricing. purchasers. they tend to be a smaller Power. a day response programs are designed with different ahead. there was some expected growth in dynamic pricing Chapter 7: Engaging Electricity Demand 155 . at estimated that its real-time pricing customers little or no incremental cost. On those days. Whether and when reduced peak demand by 17%. roughly 800 MW. and they report consider. or real-time pricing tariffs to nology across many systems now enables those their large customers. Peak-time rebate programs suffer from the same baseline calcu- industrial customers are forgoing opportunities for lation challenges described above for wholesale significant cost-effective peak load management. they are poor demand is expected to be exceptionally high substitutes for dynamic pricing programs in relative to available supply. rather Many utilities with limited penetration of dynamic than being charged a premium for consump. Why the focus on residential customers in policy discussions? Until recently. Yet these programs can be quite customers. the high cost None of these dynamic pricing plans yet of required metering precluded extension of accounts for significant peak load reduction price-mediated demand response programs participation.retail prices change hourly (or more often) to Many utilities with limited penetration of reflect actual variations in the system’s marginal dynamic pricing plans among their larger energy cost. primarily larger customers within dynamic pricing for commercial and industrial those sectors. Duke residential users. For example. there may be considerable rebates” operate similarly to critical peak gains to be realized from further penetration. “Peak-time residential sector. their administratively determined baseline during designated critical peak hours. pricing programs. pricing plans among their larger commercial and tion during those hours. perhaps in addi- peak demand management. estimated at 2010 US$750–$5. they this regard. a small number of days on which objectives and constraints. though that may in large part reflect changes in the survey methodology. and the Tennessee Valley Authority— fraction of potential savings for large electricity have long offered variants of critical peak. At the other end of the spectrum commercial and industrial customers are are plans that layer time-of-use tariffs with an forgoing opportunities for significant cost- infrequently invoked price change for so-called effective peak load management. among large given a credit for reducing consumption below commercial and industrial customers.35 programs.000 cost effective. particularly dynamic pricing. utilities to offer residential customers at least able success in both customer satisfaction and one dynamic pricing option. “Critical peak pricing” interruptible load and emergency demand programs enable utilities to designate. market-administered demand response programs. utilities will do so is an open question. Because critical peak periods.36 The deployment of AMI tech- hourly day-ahead. Our report follows recent policy charge a price for electricity consumed during discussions in focusing on extension of demand peak hours that is several times higher than the response. While metering costs to support customers. FERC’s during the early 2000s. several utilities in per customer connection. are higher than for the Southeast—including Georgia Power. to the usual time-of-use peak rate. Enrollment is comprised almost to most residential and smaller commercial entirely of commercial and industrial customers. roughly equivalent to 2010 survey recorded some retrenchment from eliminating one large baseload power plant existing time-of-use and dynamic pricing from the necessary dispatch. Georgia Power tion to existing non-dynamic rate structures.

For example. paying little attention to modeling response to supply conditions and may enable the determinants of participation or the improved energy efficiency and conservation. tion and service requirements. and predictability of expected in programs designed to reduce consumption load response. certainty or distribution of load response. even though the benefits Deployment of AMI and complementary technologies could be substantial for systems anticipating inadequate peak reserve margins. Demand operators and utilities already calculate. many regions now experiences of utilities that do implement operate with substantial reserve margins above dynamic pricing can provide insights on a host expected peak demand as a result of past of important questions about customer educa.3 PREDICTING THE BENEFITS OF conditional on participation in a given INCREASED DEMAND ENGAGEMENT program. near-term benefits for systems that expect this situation to persist. the anticipated Demand response mediated through magnitude in MW of the response generated by the program. or changing the forward element to compute. but the task poses considerable challenges. expected uptake or population decline. It reflects the tools—say. economic contraction. programs reported in plans for establishing value is likely to vary across systems as well as future demand response programs.37 The over time. and industrial customers. enhances potential demand response to supply conditions and may enable improved energy efficiency Measuring the expected impact of demand and conservation. Altering economic value of AMI-related investments— program design can compound the challenges. tially across service areas and among program designs. Estimating the expected magnitude of these Some Recovery Act–funded pilot programs are benefits may be critical to determining the now investigating these questions. participation and response. response involves predicting customer partici- pation rates as well as customer response 7. challenging to value. Analyses of demand response typically have Deployment of AMI and complementary focused on mean load changes across enrolled technologies enhances potential demand participants. We turn next to existing at peak times is likely to generate minimal evidence on those questions.38 Increased enrollment of optional plans. Each factor is likely to vary substan. Expanding the target population to add residential customers The expected economic value of a MW of to a program populated by large commercial demand response is probably the most straight. which system and data to predict responses. This 156 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . For example. from direct load control to dynamic avoided cost of supply or shadow cost of supply pricing—may require new experimentation constraints at a given moment. and the certainty of that dynamic pricing can be particularly response. changing a direct load control program’s objectives from infrequent emer- Enhancing Demand Response gency or peak-shaving calls to a more frequent balancing role may make customer participa- The aggregate benefits of introducing a new tion and response results from historical demand response program depend on the value program data uninformative about future of a given MW change in demand in response to a set of system conditions. capacity investments. Both may be difficult to assess.

Simulation system operators lack accurate prediction analyses that have focused on system-wide models for when. at least three s 2EAL TIME PRICING SCHEMES WITH PRICES THAT sources of information may help to narrow vary hourly or more frequently could reduce the uncertainty in the benefits calculus: peak-period electricity consumption substan- tially under plausible demand elasticity s 3IMULATION METHODS.response mediated through dynamic pricing magnitude of potential gains from broad can be particularly challenging to value if deployment of dynamic pricing. where.v broadly consistent results: vi In light of these challenges. and how much load deployment of real-time pricing yield some will respond.

plants. this reduces assess the general plausibility of a stylized necessary investment in peaking generating economic case and determine which param. In the long run.39 s 0ILOT PROGRAMS AND DEMONSTRATION PROJECTS. leading to lower average electricity eters or assumptions are most critical. WHICH ARE USEFUL TO assumptions. costs than would otherwise occur.

putting less than a third of load on real-time s %XPERIENCE FROM LARGE SCALE DEPLOYMENTS. penetration across customers: in simulations. s ! SUBSTANTIAL FRACTION OF THE POTENTIAL which can provide feedback on parameters of economic benefits of universal real-time customer response and impacts of variants in pricing may be secured without universal program design.

41 This analysis is closely related to the classic cobweb model in economics.40 Simulation Analysis s $YNAMIC PRICING SCHEMES DO NOT NECESSARILY reduce overall electricity consumption.42 Forecasting load with reactive demand is difficult but soluble. which tends to encourage v Some have expressed concern that this challenge may be so great as to threaten power system stability. and the focus is on high-level aggregate returns rather than analysis of sensitivity to parameter values or modeling assumptions. in voluntary dynamic pricing schemes. longer time average cost-based flat rates will realize much periods. as well as wholesale price-mediated. Chapter 7: Engaging Electricity Demand 157 . the impact of price changes.43 This paper probably is best viewed as providing a framework for analysis rather than simulation results for net returns to AMI investment. pricing can yield roughly half the potential which can help refine estimates of customer total benefits. recent work illustrates the potentially destabilizing effect of price-responsive demand in a model of an ISO that uses only data on loads in past periods to forecast future load and sets price so that generation equals forecast load. vi A recent Institute for Electric Efficiency white paper simulates lifetime costs and benefits of AMI- facilitated demand response for four “prototype” utilities. and its own past forecast errors. as a number of parameters are varied in ways that do not nest across prototypes. indicating an issue of “free riding” deviate from forecasts. provide some insights on the source and likely off-peak prices fall. Customers remaining on response over larger populations. as suggested by considerable experience internationally as well as in the U. and dynamic pricing programs. which has little explanatory power in most settings because actors are more sophisticated forecasters than assumed in the model. They also provide data on the actual creates without adjusting their consumption costs of implementation and how those patterns. ignoring past prices. customer time-of-use rates.S. For example. of the cost savings that real-time pricing ment. While Theoretical analyses of dynamic pricing can peak prices increase relative to flat rates. with dispatch under customer-directed responses to emergency load-management. and in the context of actual deploy.

brattle. Faruqui.. August 24. on knowing realistic parameter values. Unfortunately. “The Tao of the Smart Grid. or better s 2EAL TIME PRICING EXACERBATES BILL VOLATILITY still.45 from most pilots typically have been inacces- sible to outside researchers. programs implemented at large scale. a large number of small-scale customers whose demand is correlated with and a few large-scale pilot programs have system loads (e. greater off-peak consumption as well as pricing on electricity consumption.” presentation to the Michigan Smart Grid Collaborative. Figure 7. data of real-time pricing policies. those who use more air experimented with various forms of dynamic conditioning at system peaks).pdf. Figure 7. With their load-shifting. these simulation results are useful to draw general conclusions. A substantial fraction of the annual electricity charges accrue during a very few. but very Pilot Programs high-price peak hours. making it difficult In general. Source: A. Additional data available hedging strategies that can manage will be collected over the near term as system- or reduce bill volatility is likely to play a wide AMI deployments supported by the significant role in encouraging the adoption Recovery Act come on line. Making pricing at the residential level. Tech=technology.5 is for assessing the qualitative impacts of dynamic a summary presented by researchers at the 158 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .44 theoretical studies cannot replace empirical evidence obtained from pilot projects. 2011. CPP=critical peak pricing. PTR=peak-time rebate. RTP=real-time pricing.g.5 Peak Load Reduction from Dynamic Pricing Pilot Programs by Rate Design and Technology 60% TOU TOU w/Tech PTR PTR w/Tech CPP CPP w/Tech RTP RTP w/Tech 50% 40% Peak Reduction 30% 20% 10% 0% 1 109 Pricing Test Note: TOU=time-of-use pricing. MI. http://www. If off-peak consumption many simplifying assumptions and dependence increases exceed peak-period declines.com/_documents/UploadLibrary/Upload973. Lansing. suggesting substantial swings in monthly bills—especially for In the last decade. aggregate energy use will rise. however.

Brattle Group based on their proprietary data of the business or policy case for AMI invest- on 24 pilot projects implementing 109 variants ment.49 technology variant and the height of the bars represent the peak load reduction observed. Moreover. pilot project outcomes must be that indicate high-price periods. additional data or econometric modeling of price and technology interventions. such as programmable controllable thermostats. The results are to the strength of individual pilot design and arranged in ascending order within each implementation quality. or home-area interpreted cautiously with attention given energy management systems. enabling technology. energy lights or “orbs” Moreover. Several caveats should category of rate design and technology.46 In this may be necessary to improve the precision of figure. pilot project outcomes must be interpreted The results are grouped by rate design and cautiously with attention given to the strength of whether the program provided customers with individual pilot design and implementation quality. each bar represents a different price and expected mean reductions. The be highlighted: results suggest consistent evidence that consumers respond to dynamic prices by s -OST SIGNIlCANTLY.

the pene. broader overall customer group. drop from the study and revert to standard Extrapolating these results to make specific rates. in the California Statewide demand response is an important determinant Pricing Pilot—and has not been carefully Chapter 7: Engaging Electricity Demand 159 . participants may be among the customers The estimated average peak load reductions by most willing to respond to price signals. Average peak load reductions by “treatment” (dynamic pricing) and “control” participating customers vary more (standard tariff) groups were selected from customers who voluntarily participated in than tenfold across the pilots. s 0ILOT PROGRAM PARTICIPANTS MAY BE REPRESEN- This variation is in part explained by the tative of consumers enrolling under a similar magnitude of each pilot’s price differential optional dynamic pricing tariff in a given between peak and off-peak periods. widely across pilots—for example. participating customers vary more than tenfold as a result. these results are powerful. participation rates appear to be one of the all of which involve utility-specific adjustments. in some cases. but extrapolating participa- tration of central air conditioning among the tion rates from one program type or area to customer base. Enormous response variation remains even when the results are grouped by rate design. it appears that considerable overall program benefits. suffer in varying degrees from some form of selection bias in participation. virtually all studies substantially. weather during the experiment. In most cases. may not be representative of the across the pilots reported in Figure 7. In many experiments.47 most significant unknowns in determining Nonetheless. Discretionary participation implies that quantitative predictions is more problematic. and. DESPITE EFFORTS TO IMPLE- reducing their electricity consumption during ment an experimental protocol with treat- high-price periods. quite ment and control groups.48 For situations in in Connecticut Light & Power’s pilot to 20%. another may be difficult. Determinants of and the availability of automated technologies. customers assigned to the the dynamic pricing tariff As evidence that consumers will respond (the “treatment group”) were permitted to to price signals. service territory.5. the pilot. which the quantitative magnitude of potential at least initially. from 3% study controls are employed. Participation varies heterogeneity persists even after many cross.

but we do not against an untreated control group. population. the fraction of customers attract greater consumer attention that wears selecting into the pilot. recently completed or will soon complete tion on the remaining (perhaps hotter) days. short-term novelty or attention effects. A number of utility systems have on half the critical peak days and no reduc. For example. it will be difficult to be greater. limiting the strength of though this may be an important input to statistical inferences and making it dangerous regulators considering dynamic pricing plans to generalize the results to the broader (discussed in greater detail in this chapter). bution of responses by the group facing dynamic prices. studied. Even if the selection at work in the s #ONSUMER RESPONSES TO PERMANENT DYNAMIC pilots is related to the behavior of customers pricing plans may differ from those in short- likely to participate in voluntary dynamic term pilot projects. these gaps. a System-wide deployments of AMI combined 5% peak reduction on every critical peak day with optional dynamic pricing tariffs may may have dramatically different implications provide the data needed to fill in at least some for system operation than a 10% reduction of these gaps. though this information may System-wide Deployments be important to operate the system and assess the value of demand response. that combined with optional dynamic mask a wide variance in the number of hours pricing tariffs may provide the data for which data are collected. Others do not have adequate There is some evidence to show that most control groups or baseline data that allow reduction in demand comes from a relatively analysts to measure treatment group responses small number of customers. although two studies that from pilots to those for the tariff under followed customers across two summer consideration. suggesting the first effect dominated any under optional programs. statewide legislative or regulatory mandates. Long-term responses may pricing programs. or Smart Grid Investment Grant 160 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . know whether those could be identified in advance and targeted or how persistent s -OST PILOT STUDIES REPORT RELATIVELY SIMPLE their identity is over time. rollouts of AMI technology to all or most of though both scenarios imply a 5% mean their customer base as a result of utility initia- reduction. few customers. as when consumers invest in predict responses to system-wide programs complementary technology or adopt new without information on the pilot protocol behaviors. They could be smaller if pilots selection rule.50 s 4HE QUALITY OF INFORMATION PROVIDED BY PILOTS s &EW PILOTS HAVE ANALYZED THE DISTRIBUTION OF varies widely. and correlates of off over time. Few pilots have measured customer participation rates. Future pilots and system seasons report similar or increased responsive- deployments should provide greater informa. Many pilots enrolled relatively responses or bill impacts across customers. This may make residential responses for more than a few it difficult to match predicted participation months or a year. In addition. tives.51 measures of responses. there has been little analysis of the predictability needed to fill in at least some of of the mean response or the temporal distri. usually mean reduc- tion in peak-period consumption relative System-wide deployments of AMI to a baseline or control group’s behavior. ness among those who participated in both tion and focus on participation determinants years.

At programs. equipment. could yield and policy makers to search for new types of significant social benefits. or in improving our ability to assess the potential home improvements that consumers must benefits of AMI-facilitated demand response incur to achieve the reduction in energy use. tially higher-impact programs based on infor- such access remains uncertain as of this report’s mational feedback. Facilitating data collection and ensuring data accessibility could improve the As noted earlier. however. some plan to offer customers a menu of pricing options. customer defaults and opt-in reducing energy use through programs that or opt-out rules. SmartGrid. combined with an objective of greater Smart Grid Information Clearinghouse consumption impacts. including the dynamic pricing options particularly for residential customers. have not yet committed to AMI. has prompted industry (www. an increase of 24% over to these programs are collected in a systematic 2009.4 REDUCING RESIDENTIAL ENERGY or committed to some form of dynamic pricing CONSUMPTION to follow the AMI deployment. provided these data programs.53 comparable to the comparisons across installations. design. apart from any social share data from smart grid projects. the installations are all likely to make informa.52 Ratepayers typically fund utility- refine or validate predictions for dynamic pricing run programs in which all consumers pay a impacts that come out of the pilot studies.gov (www. and subsidies for a Investment Grant funding were expected to variety of energy-enhancing home improve- collect data on consumer responses under a ments to participating customers. Unfortunately.gov) and the This. Since the 1970s. including externality associated with electricity use. costs largely offset the savings from of Energy to create mechanisms for utilities to reduced consumption. this variety could help to programs. the diversity of the pricing generation cost in many regions of the country.4 billion.smartgrid. a number of companies already design of dynamic pricing plans for utilities offer commercial or industrial customers that have deployed AMI technologies and the services that can use granular consumption quality of adoption decisions for utilities that data. Department these levels. These structured experimental protocol. and that are offered.projects.sgiclearinghouse. and availability of hedging or include a broad mix of targets that may range commitments that guarantee bills will be no from energy-efficient lightbulbs and appliances higher under dynamic pricing than under to insulation and weather stripping. Energy- historic tariffs during some specified transition efficiency budget expenditures in 2010 period (a year is common). Efforts led by the U.S. Many of these utilities have proposed 7. which recent estimates suggest ments also should collect and make available cost an average of four cents per imputed comparable data to facilitate learning and kilowatt hour (kWh) saved. If data on responses exceeded $5. such as that provided by AMI technologies. energy audits. and the ongoing.org/). many private costs of efficiency programs: the tion on consumer response of enormous value additional expenses in devices. programs are associated with modestly reduced ments in systems not subject to these require. The scale of average wholesale electricity price or avoided these programs. of which 88% was in utility-administered and thoughtful way. permanent nature of This calculation does not include. writing. schemes. Chapter 7: Engaging Electricity Demand 161 . and deploy. small surcharge on monthly bills to support the program budget. which may provide appliance Deployments made with Smart Grid rebates. consumption. many electric utilities have There is considerable heterogeneity in program been tasked with improving energy efficiency. focusing attention on potentially are made available to researchers and utilities lower-cost behavioral interventions or poten- that are not grant recipients.

The untapped potential for reducing electricity use with a range of 0. do recipients of feedback find the substantial recent projects tracked roughly information provided useful. and content of the feedback that leads to conserva- generally do not resolve whether the responses tion.57 Profiling and may lie in residential behavioral changes. or simply the incidence of messaging. although the magnitude as a smart meter or part of a research trial.1). observations. systematically. 162 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .55 makes use of mailings to customers and does not rely on technologies. Much of the untapped data collected across 11 utility pilots (750. Profiling. report program costs and quantify the net tricity consumption. These appear to presented as a routine replacement rather than reduce consumption. some results are consistent tion. which However.K. are long term or transitory. the combination of dropout challenges.9%.5 cents for 1974 and 2010.9%–2. For example. frequently suffer from further improve a program’s cost-effectiveness. may lead to unrepresentative or broadly raises the question of whether it is the non-random samples of participants. control groups reports average consumption reductions of roughly 2% across programs. a recent analysis of a large set of energy consumption. assessment. perhaps even more than those individuals most likely to respond. even these experiments with messaging being more important than confronted significant participant selection and content.K. for a broad range of inter. may demand response pilots.000 potential for reducing electricity use may lie in households) that includes pre-treatment residential behavioral changes and modifica. targeting programs to high-consumption households may increase the impact to around Many experimental programs over the last 6%. randomized treatment.000 customers across treatment and control of providing feedback serve as a reminder groups in the U. with of programs on different continents between program costs running roughly 2. measurement.”60 of the effect tends to be consistently modest. or does the act 60.58 This estimate is based on a feedback pilots using in-home displays reported vendor-run indirect feedback campaign that reductions averaging 7%.56 AMI and real-time display treatment was associated with reduced consumption for all Feedback-based behavioral programs that the previously mention U. of conservation goals?59 While this causal ventions (see Box 7. One of the very few experiments to informational feedback on household elec. however. relative to pretreatment consumption several decades have assessed the impact of levels. One of the more That is. to suggest cost-effective modifications to reduce For example. energy-efficiency and conservation targets behavioral intervention programs at field experiments. and design flaws that sharply limit the generalizability of their The experience with feedback programs more findings. which the final report a positive norm of reducing electricity speculated “may be related to the fact that the consumption have been the subject of increased meter replacement [for Scottish Power] was experimentation in recent years. and tions to traditional consumption patterns. Despite considerable difference has not generally been tested attention to program design and implementa.54 A recent comparison of direct each kWh saved. Average energy reductions economic impact suggests that behavioral in kWh ranged from 4% to 14% across a variety interventions may be cost-effective. trials except provide a “social nudge” to consumers based on Scottish Power’s trial.

This is also an option for consumers who have access to AMI meters but for whom the cost-effectiveness of investment in additional energy management technologies is unclear. and advanced metering. the FINDING design of behavioral interventions to reduce While a variety of programs suggest electricity consumption seems a promising area that energy savings are possible through for continued experimentation and research. noteworthy because of its large scale. and breadth of technologies and inter- AMI installers appears to have influenced ventions. Moreover. the advanced metering generally were more Office of the Gas and Electricity Markets (Ofgem) effective than those without one or both of Energy Demand Research Project in the U. such programs can be designed and imple- mented at relatively low costs.61 Since 2007. Initial experiments suggest that merits further experimentation. the design of This seems especially true of interventions that cost-effective behavioral interventions draw on fairly simple and inexpensive informa- tional feedback. detailed billing. raising the possibility residential customers of four different energy that behavioral interventions may be as suppliers participated in a variety of trials significant as technology.1 THE OFGEM ENERGY DEMAND Highlights of the findings include the following: RESEARCH PROJECT r 1  SPHSBNTGPDVTFEPOFOHBHJOHBOENPUJWBU- Of the many electricity systems that have ing consumers to reduce energy consumption undertaken pilot projects to evaluate the effects with a combination of in-home displays and of informational feedback to customers. Technical and management lessons from the project are informing the planned nationwide deployment of electricity and gas meters to enable the provision of informational feedback to consumers. BOX 7. r 8  IJMFNBOZPGUIFJOUFSWFOUJPOTSFDPSE incentive programs designed to reduce or shift consumption declines generally in the range consumption. there was limited consis- sity research centers assisted with independent tency in energy savings across energy statistical analysis and program design.000 consumer responses. they could be a suitable near-term alternative to AMI-mediated technologies for systems that have not yet committed to broad AMI deploy- ment. project report was prepared by consulting firm r $  POTVNFSTGPVOEQSJDFBOEDPTUEBUBNPSF AECOM. (roughly one quarter as controls). Univer. more than 60. is these technologies. Chapter 7: Engaging Electricity Demand 163 . the final suppliers testing similar interventions. including tests of in-home display units. long time r 5 IFOBUVSFBOERVBMJUZPGFOHBHFNFOUXJUI frame. informational feedback. of previous studies.62 useful than quantity or carbon emissions data.K. Given the current state of knowledge.

outreach. providing maximum bill guarantees for a specified term following take-up of a dynamic For utilities that are deploying AMI throughout pricing plan. and potentially substantial benefits. Regulators must be prepared to customer resistance to change. at least tures at the same time new meters are installed as an option. and that Implementing rate hikes or new price struc- dynamic pricing will be widely available. we believe that by 2030. or will reduce their average energy acquisition efficiency programs. How individual systems will or may exacerbate consumer resistance not only should move toward that situation over time is to the tariffs.5 EXPANDING DEMAND ENGAGEMENT: effective programs to engage customers and FINDINGS fund periods of higher customer service costs through the transition. tariff offerings available to customers. and service. Dynamic pricing the development and penetration of dynamic plans are likely to be embedded in a menu of pricing options and their associated benefits. and support transitions to dynamic Competition in electricity retailing may pricing that include extensive customer educa. As noted earlier.65 much less clear. dynamic pricing programs likely to gain if best practices for transitions promise potentially substantial benefits. This is highlighted in a recent energy. not fall. or study of 21 organizations with successful direct load control programs.S. mitigate potential adverse reactions include running AMI side-by-side with existing meters For utilities that are deploying AMI throughout for some time. AMI will be widely deployed across the U. A monopoly utility has suggest that customer service demands may the same opportunities but may have weaker actually rise. Inducing provided that appropriate default service tariffs general acceptance of dynamic pricing will are established or the monopolies that provide require a different approach toward customer distribution services (so-called wires firms) are engagement than the one most utilities have excluded entirely from the business of selling taken in the past. complement AMI deployment in hastening tion. 7. which identifies key themes and wholesale prices—through dynamic pricing best practices that emerged from its in-depth tariffs. It needs the installation of new meters. Strategies utilities have adopted to committed to universal deployment of AMI. utilities and approve investments necessary to develop regulators may have strong incentives to be 164 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . energy management technologies. dynamic pricing. We focus on these first Electric’s smart meter installations demon- in the context of systems that have already strate.66 Retailers that can successfully reduce report released by the Smart Grid Consumer their customers’ usage during periods with high Collaborative. in the months following incentives to innovate (see Chapter 8). some ongoing consumer protests over Pacific Gas & generalizations arise. if publicly owned) service expenses ultimately decrease with AMI in an inherently political environment. but also to the technology itself. stage. realize this potential. Although no one-size-fits-all Addressing this after the fact can be costly. Given installation. though most to obtain approval for new rate structures from business cases assume call center and customer its regulator (or equivalent. some pilots costs and consequently can offer their customers and early system deployment experiences lower average bills. Utilities and their customers are their systems. for example— customer-facing AMI. utilities must carefully plan. as recommendation can be formulated. dynamic pricing programs promise on actual consumption data for the period.63 To are analyzed and widely shared. offering bill comparisons of charges under current and future tariffs based their systems.64 Moreover.

71 several pilots suggests that low-income customers respond to higher prices.68 by confronting that resistance. offsetting part or all of this effect consumers. such as those with peak demands be at least initially. as we expect it will generally the system. Retail competition conse- quently may encourage greater experimenta. while few pilots fraction of consumers may reduce system-wide have focused specifically on low-income energy costs.69 to win customers. their demand responsiveness in as more equitable than flat-rate tariffs because they terms of peak load reduction was below the allocate more of the cost of consumption to those who mean for all customers in the program. have strong incentives to may be inherently less costly to serve and may find ways to use the benefits of dynamic pricing benefit from dynamic pricing. wide basis is not free. Peak-time rebates may where implemented appropriately. Moreover. in the PowerCents program coupled a basic time-of-use rate with rebates for peak-hour FINDING consumption reductions relative to baseline on Well-designed and well-executed preannounced critical peak days. adjust peak consumption. customers—especially residential customers tial and smaller commercial customers—subsi. —to enroll in time-varying tariffs. ways to limit distributional impacts tion than most regulated monopolies will exist. consumers who do not participate in peak- Many low income consumers may have flatter period reductions eventually may experience than average consumption profiles. baseline consumption to increase their poten- tial rebate and tend to be associated with lower mean responses than critical peak pricing plans. particularly for lower-income also must be high enough to cover the revenue households. one recent analysis of data from on base rates. Dynamic pricing plans may eventually come to be viewed as more equitable than flat-rate While regulators in some jurisdictions have tariffs because they allocate more of the cost agreed to make time-of-use or dynamic pricing of consumption to those who are responsible plans a required or default contract for large for imposing those costs on the system. Concern about distributional impacts of as shown in Figure 7. For example. To customer backlash to AMI technology and be clear. Moreover.cautious and find relatively little to be gained that are coincident with system peaks.5. This provides a form of insurance: customers gain from customer engagement programs and tariff reducing peak electricity use but do not pay transition policies are needed to avoid higher rates if consumption is unchanged. suggesting that This should not be seen as insurmountable.67 are responsible for imposing those costs on the system. there is constant rate structure applied to most retail widespread reluctance to require smaller electricity customers—and virtually all residen. But all else a benefit from dynamic pricing without any may not be equal: demand response by even a demand response. providing this insurance on a system- dynamic pricing plans. may impede transitions needed to pay for rebates to customers who to dynamic pricing under monopoly retailing. the available evidence Chapter 7: Engaging Electricity Demand 165 .70 The level of base rates dynamic pricing. may provide incentives for customers to raise facilitate dynamic pricing. Competitive Customers with less peaked demand profiles retailers. all else equal. though Dynamic pricing plans may eventually come to be viewed in most cases. suggesting higher average bills. Retail competition. the peak rebate tariff tested undertake. If participa- dizes customers who impose higher costs on tion is voluntary. in contrast. The commercial or industrial customers.

But consumer participation rates in demand it seems unlikely that many consumers will response programs have not been well-studied. there is a chicken–egg issue here: options. and utilities’ price 166 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .73 One of the solution is likely to involve assignment of the most consistent findings across pricing pilots responsibility for technology installation is the much greater responsiveness of demand decisions. While decide to adjust consumption accordingly. automatically based on preassigned instruc- tions. Implementing complex dynamic FINDING pricing schemes before low-cost Concern about the impact of higher or more enabling technologies are available volatile bills. solid-state meters. ownership. and doing so customer choice into dynamic pricing plans. Or consumers Capturing significant benefits from dynamic could opt for more complex and expensive pricing is likely to require investments in home energy management systems that inte- complementary technologies. and payment to partici- in the presence of enabling technologies. particularly for low-income may limit realized benefits and consumers. without dynamic pricing.72 Experimentation may be technologies are available may limit realized needed to determine how best to design benefits and cost-effectiveness. tion than regulated monopoly. particularly when the less than universal. enabling technologies offer few benefits. which can retailers to be active in this market if policies do receive price or other signals from the distribu. This pating customers. and other major loads and optimize those that permit customers to automate use based on granular consumption data from their responses to price changes.74 Implementing complex dynamic much lower participation rates than default. or an energy “orb” that signals high price FINDING periods by changing colors. in be unnecessarily expensive even with dynamic critical peak pricing programs. if present. and accordingly. We would expect third-party could be something as simple as a program. not restrict their activity or embed enabling tion system and adjust air conditioner load technology investment in utility ratebases. Consumers may be able to respond manually in Attempting to install enabling technology limited ways and at relatively low cost to some across a utility’s entire customer base is likely to forms of dynamic pricing. suggests that opt-in plans generally lead to signals. service providers or. high-cost consumption. For example. HVAC systems.75 The most cost-effective overall bill impact may be small. there will be little incentive to develop or deploy them. particularly grate “smart” automated appliances. retail competition may provide increase because consumers cannot easily avoid stronger encouragement to such experimenta. competitive mable controllable thermostat. especially if enrollment in told the day before about price spikes and can dynamic pricing plans is not mandatory. may limit the perceived political cost-effectiveness viability of dynamic pricing but this can be addressed by well-designed menus of rate Of course. customers’ price and comfort preferences. plans. consumers are pricing tariffs. may also lead to consumer frustration as bills As argued here. want to spend a lot of time and effort managing most analysts agree that enrollment will be far their electricity usage. pricing schemes before low-cost enabling or opt-out.

a third-party system efficiency persist or intensify. for some utilities. Many customers may choose not to deployment at scale and the expected participate in demand response programs. information on benefit of expanding residential price-responsive many of these is likely to improve considerably demand may not be not clear. For these customers.Finally. systems may be reliant on demand ment over a few years. and plug-in hybrid vehicles controllable thermostat. Current at least for some time. information tech- nology. there is considerable implement price-mediated demand response uncertainty about both the cost of AMI and programs. Where proposals because of the lag between submis- operational benefits are low relative to deploy. and the participation rates and system benefits are precise magnitude of demand response benefits likely to improve. the economic value of expenditures by the distribution utility in meter early replacement of existing meters with data management systems. Fortunately. short of the cost of deployment. Their net customer service. and subsequent staged deploy- ment costs. or. particularly over the near term. Whether those costs become more important and concerns for are paid directly by the customer. the distribution utility itself. sion. as systems because. pure area network or as simple as a programmable electric vehicles. approval. the deployments are likely to improve the full metering cost is incurred with no prospect quality of information on costs and benefits of demand response benefits as offsets. tion technology.6 CONCLUSIONS AND installed meter over the past five years. FINDING Using AMI to enable real-time pricing or other For systems with limited operational demand response programs requires significant savings from AMI. and customer education and service. For these systems. that investment would produce.76 We confidently expect the demand for elec- Moreover. some cases. Utility and thus enable better decision-making. though RECOMMENDATIONS more recent industry estimates suggest a possible decline to $150–$250 per meter. such as those surveyed in Chapter 6. Similarly. the belong in the economic cost–benefit pure operational benefits of AMI may fall far calculation. This may be especially likely for systems that have recently Actual costs of deployment may differ from installed automated meter reading (non-AMI) the prospective estimates submitted in utility systems throughout their customer base. Chapter 7: Engaging Electricity Demand 167 . a competitive energy retailer. they As noted in Chapter 6. The value of making demand complex and expensive as a separate home- more responsive will increase as VERs. in yet committed to a system deployment of AMI. at present. Predictions may be response and conservation benefits to make a especially uncertain for the costs of informa- business case for investment. and AMI investments are not yet sunk. to an important extent likely to incur additional investment costs for because of increased adoption of dynamic enabling technology—which may be as pricing over time. customer education. rate filings. we turn to those utilities that have not aggregator. suggest that the expected all-in cost of AMI deployments averaged perhaps $300 per 7. estimates of demand response its required associated technologies. customers choosing to participate tricity to become increasingly responsive to in voluntary demand response programs are system conditions. demand response benefits. in new AMI meters is uncertain due to limited addition to meter acquisition and installation information on both the “all-in” costs of costs.

If and we would expect those commitments to be dynamic pricing tariffs are offered but not made. across service areas to provide a synthetic control group. This structure also can improve Early adopters confront a number of challenges a utility’s ability to allocate customer service and responsibilities. the largest cost of to universal AMI deployment throughout their enabling greater demand engagement via systems. reduced meter-reading labor. new technologies. AMI if their information technology systems have investments may pass cost–benefit tests without simultaneously been upgraded to accommo- requiring any significant demand-side benefit. utilities life. particularly plished that objective. automated residential appliances. and potential new communications and ently committed to widespread AMI “behind-the-meter” technologies is essential deployment. and the direction of innova- ting fine-grained. improving rapidly. Ensuring that new investment and essential for truly dynamic meters and any associated equipment installed pricing. date the data flow from AMI systems. as discussed in greater a system in which some form of time-varying detail in Chapter 9. partic. 168 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . manually read. In addition. customers and of both automating and consumers are concerned about the privacy measuring their responses. in fact. immature and costly.77 These systems may find the economics should collect baseline consumption data of near-term AMI investments attractive and before exposing consumers to new rate designs join the utilities that have already committed and may stagger the rollout of new rate options to AMI. But. as seems likely at least initially. real-time price signals to tion is difficult to predict. dynamic pricing has been sunk. and Structuring demand response program imple- theft reduction largely offset the capital costs of mentation using best-practice pilot design will universal AMI deployment across their service enhance the ability of researchers to measure area over an estimated 20-year useful meter programmatic impact. For example. to stages of evolution. Capabilities and costs are ularly AMI. Early AMI Adopters For the many utilities that have already installed AMI technology across their systems or are Many utilities have already made commitments committed to doing so. systems that operate mandated. reduce subsidies to flat-rate customers and high customer turnover or high billing losses encourage efficient evolution toward universal may find that the operational cost savings from dynamic pricing defaults over time. and unforeseen problems that develop in other complementary technologies are relatively implementation. our recommendation for how to by the utility offer interoperability with current proceed depends on whether a utility is pres. For other utilities. Stranding consumer or prices becomes the default for many if not all utility investment in meter or control technolo- customers. are reducing the cost of transmit. gies that are not compatible with later genera- Because AMI deployment is both a major tions could be quite costly. is likely to differ across utilities. remote connection and disconnection. outage detection. traditional electromechanical default service tariffs should be designed to meters and those with low customer density. some have already accom. the path from and security implications of AMI. Complementing this. Home energy management resources and respond effectively to any systems. The system today’s flat-rate pricing system for all but the standards for cybersecurity are in the early largest commercial and industrial customers. For example. to preserving the value of these investments.

The capability-adjusted costs of AMI investments in customer education and and complementary technologies seem likely to decline as innovation advances and manu- engagement programs to encourage facturers gain greater experience and scale in migration to universal dynamic pricing. costs and demand-side reconsider next year. This is likely to be the case for many of Grants were designed to have a crucial informa. Deferring deployment decisions also reduces the chance that a utility will be locked into an early but inferior technology. Early adoptions have tion over the next few years. operational information gained from innovative programs benefits of universal AMI deployment may and investments is critical to efficient evolution offset only 50%–60% of the system installation at the national level.These requirements should be viewed not as Systems without Present AMI Commitments impediments but as requisites to effective implementation.org/).” but rather should focus on interoperability and “now or reconsider next year.sgiclearinghouse. Utilities in this R E CO M M E N D AT I O N situation may rationally decide not to commit Where a commitment to AMI technology investment dollars to AMI until some uncer- tainty is resolved.” but rather “now or are critical to realizing the greatest value from tain implementation early AMI adoption. Ensuring timely and they may have many years of useful life reporting to these repositories. Smart Grid Investment costs. Chapter 7: Engaging Electricity Demand 169 . Given the volume of existing AMI an ultimate goal of substantial real-time installations and those committed to comple- pricing penetration. sharing For many other U. the quality of an important research and demonstration information on deployment costs and the component: information on investment range of demand-side benefits also is likely to costs and the results of demand response improve considerably in the near term. AMI investment decisions are not pants and researchers outside the utility sector depending on uncer- “now or never. These meters capture a for sharing data via www. and to substantially positive making data accessible to both industry partici. using them to remaining.smartgrid.S.78 Crucially. utilities. is collected and then revisit the decision. AMI investment is already in place.gov and significant fraction of the reduced meter- the Smart Grid Information Clearinghouse reading costs promised by AMI technologies. Utilities that have not yet committed to AMI rollout programs should accordingly be collected have the option to wait while this information and shared widely. and the Department non-AMI meters now deployed in the U. As in many other areas. with production. in net present value. that of Energy has led efforts to create mechanisms can be read remotely. investment decisions decisions are not “now or never. (www. In this case.” benefits. the utilities operating the roughly 50 million tion-sharing component.S. Waiting until lessons are learned from the first Utilities should design and follow a wave of AMI installations may yield myriad transition path that includes appropriate benefits.” compatibility with later generations in meters and associated technologies. investments in AMI collect similar information for deployments that might range from significantly uneconomic were not supported by stimulus funding.

if not all. We expect by the end of our study charging may offer additional opportunities for horizon in 2030.79 170 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . should recognize the option value of tioning and electric hot water heating have waiting to learn from early deployment data substantial current enrollments in some utility service areas. as noted earlier. utilities direct load control in regions with significant will have rationally deployed AMI technologies. could be expanded on both the intensive and demand response program design and margin—increasing the frequency or set of benefits. R E CO M M E N D AT I O N tion mode must forego residential demand The decision whether to adopt AMI need response. This does not imply that utilities in this evalua. customer responses. electric vehicle penetration. even where AMI is installed and especially where it is not. U. For example. we are advocating care. Electric vehicle nite delay. not indefi- using direct load control. Direct load control programs on costs. technologies. circumstances under which utilities can invoke load control—and the extensive margin— expanding the set of utilities and customers To be clear.S. complement to dynamic pricing for many customers. These programs do and we hope that policy makers will have not require universal AMI for implementation facilitated movement far along the path to and may be a highly cost-effective alternative or universal dynamic pricing. most. direct not be made immediately. Decision makers load control programs targeting air condi.

C. “A Review of Competitive Electricity Markets (Berkeley. Schwartz. 4 D.” press 2 For example. no. no. 4 (2010).” Resource and Energy Economics of Achievable Potential from Energy Efficiency 33. 19 Federal Energy Regulatory Commission. “Demand DC: Edison Foundation Institute for Electric Response in U. DC: American Council for an Energy-Efficient Economy. Contract with Southern California Edison.” Electricity Journal 22. 25 Neenan Associates. A. “The Power 18 Federal Energy Regulatory Commission. “EnerNOC Signs EfficiencySMART™ 12 Management (Boston. Assessment of Demand Response and Advanced Metering Staff Report (Washington.pdf. no. Utility-Scale Smart Meter Deployments. no. 21. see note 6 of Dynamic Pricing. (2009): 42–56. 23 Federal Energy Regulatory Commission. Allcott. “Rethinking Real-Time 22 Electric Power Research Institute. 7 D. Faruqui and S. MA.” Journal of Environmental Psychology 25. http://www. Neumann. 2005). CA: Intervention Studies Aimed at Household Energy Lawrence Berkeley National Laboratory. Faruqui. NY ISO Price Responsive Load Program Evaluation (Syracuse. 4 (2011): 820–842. see note 6 “Behavior and Energy Policy. 8 For example. and J.S. Bureau of Labor Statistics. State of the Markets Report 2009 27 K. research report U052 (Washington. Sergici. white paper (Washington. 9–10 (2011): 1082–1095. (San Francisco. “Household Response to Dynamic Pricing of 13 North American Electric Reliability Corporation. 5970 above. 2010). 52. AK. Emerging Flexible Resources (Washington. and A. 193–225. Sioshansi. edisonfoundation.” Public Utilities Fortnightly.net/iee/issueBriefs/SmartMeter_ Special Report: Potential Reliability Impacts of Rollouts_0910. see H. Vlek. NY. Expenditure Survey (Washington. March 23. Abrahamse. Advanced Metering Initiatives and Residential Feedback Programs: A Meta-Review for Household Electricity-Saving Opportunities (Washington. CA. 3 U.S. Wood. Leeds. see note 6 6 Federal Energy Regulatory Commission. see note 6 15 Foundation Institute for Electric Efficiency. tion. 16 North American Electric Reliability Corporation. 24 G. and ISO/RTO Council. 20 Market Monitoring Unit Report (Little Rock.” Energy 35. above. 2011). 2010 above. and Kathan.org/publications/2010- between Demand Response and Energy Efficiency: State-of-the-Market-Report. Kathan. Edison Federal Energy Regulatory Commission. 3 (2005): 273–291. and T. 38. Goldman. Electricity Markets: Empirical Efficiency. 2010). and J. Chapter 7: Engaging Electricity Demand 171 . York and M. technical report 1016987 9 H. (March 5. 2009). Allcott and S. DC. 2 (2010): (DADS): Phase I and II Final Report (Washington. and J. page 29. Assessment Electricity Pricing.” Science 327.pdf. Cappers. 2009).” Demand Response Availability Data System Journal of Regulatory Economics. Hledik. Exploring the Relationship 2011). 34. see note 6 Council for an Energy-Efficient Economy. Ehrhardt-Martinez. Heffner. 10 Electric Power Research Institute. Rothengatter. Primer on Demand-Side EnerNOC. DC. 2010): 1204–1205. 1 Charles River Associates. C. 5 A. Electricity: A Survey of 15 Pricing Experiments. DC. http://www. (2010–2030).spp. 2011. Residential Customer. DC. (2009). 2010). Allcott. page 34. Conservation. Donnelly. Residential Electricity Use Feedback: A Research Synthesis and Cappers. Plans & Proposals (Washington. “The Missing Link. page 39. 1526–1535. Evidence. no. and W. US Smart Grid Forecast 2010–2015 March 2007. Configuring Loads as a Resource for L. 26 Economic Framework (Palo Alto. CA. F. Kushler. 2009 Consumer 14 S. Yee. see A. 2002). 3 above.S. no. no. 2005). see note 6 above. and Demand Response Programs in the U. A Review of Experience and Discussion of Key Issues. Mullainathan. Vojdani. CA: GTM Research. K. Laitner. DC: American 21 Federal Energy Regulatory Commission. 2011). 2009). 17 P. DC. The Costs and Benefits of Smart Meters for 2010). above. Faruqui. R. and D. 2011).REFERENCES 11 Federal Energy Regulatory Commission. H. see note 17 above.” Journal of Public Economics 95. and G. “Social Norms and Energy Conserva. Southwest Power Pool. 2010 State of the Market. Tsoukalis. Steg. Goldman. L. release. (Palo Alto. 2002).

Summary of Rate-Payer 52 Run Effects of Time-Varying Prices in Competitive Funded Electric Efficiency Impacts. Smart Grid Market Forecast 2010– A.” The Electricity Journal 23. Puller (Chicago: University of Chicago 33 J. Bushnell. Budgets and Electricity Markets. Atlanta. Foundation Institute for Electric Efficiency. (2009). 2011. see note 1 above. Borenstein. “Rethinking 47 2015 (San Francisco. Journal of Economics 52.pdf. and F. 2005). Prices. 44 Incentives for Efficient Demand Response. Household Response the Costs and Benefits of the Smart Grid (Palo Alto. “Surprising Deployment Developments. W. “On the 41 Stability of Wholesale Electricity Markets under 29 EnerNOC. to Dynamic Pricing of Electricity: A Survey of CA. 2011). Sergici. 127–155. Mitter.ieee. Chao.com/_documents/UploadLibrary/ 36 G. Faruqui. Faruqui and S. Smart Grid Upload973. 4 Expenditures. Faruqui. see note 2 above. M. presentation to the Michigan Smart Grid 35 Charles River Associates. see note 40 above. 1 (2011): no. A. 2011). under Real-Time Pricing. no. 2009).187 N. P. Roozbehani.” in Electricity No. EL09-68-000 (2009). and S. see note 40 above. Collaborative. 30 Comverge. 34 Federal Energy Regulatory Commission. above. Hledik. white paper (Washington. “The Cobweb Theorem. Faruqui and Sergici. Francisco. see note 47 above. and Faruqui. prepared for the Electric Power Supply Association. F. Wood. 28 Ibid.” IEEE Smart Grid Newsletter. 134 FERC ¶ 61. D. 2011). DC: Edison Residential SmartRate: Ex Post Load Impacts (San Foundation Institute for Electric Efficiency. Sergici. Vendor Ecosystem (The Cleantech Group.” Energy Journal 26. Faruqui and S. M. “The Tao of the Smart Grid. CA: 2010). 2010). “Providing Holland and Mansur. Sergici. T. Schwartz. 2010). http:// www. R. U. Cooper and L. Griffin and S. January 2010.” Public Utilities Fortnightly. brief (Washington. Leeds. and Faruqui and Sergici. and S. Borenstein. 1 (2010): 7–20. CA: The Brattle Group.S. M. 9–18. Wood. page 34. “The Long-Run Efficiency of Baltimore Gas and Electric Company Experiment. (Boston. 2011). http://smartgrid. December 15–17. 2011. Hobbs. 40 A. and Sergici. 50 2010 Long-Term Reliability Assessment “Dynamic Pricing of Electricity in the Mid- (Washington. Wolak. “When Press. Atlantic Region: Econometric Results from the 39 S. 49 deployment-developments. 43 no. A. no. Sullivan & Co. Deregulation: Choices and Challenges. 48 August 2011.” Real-Time Electricity Pricing. no. For example. 2 (February 1938): 255–280. see note 6 A. Wood. and P. no. “Time-Varying Retail Electricity 45 Federal Energy Regulatory Commission Docket Prices: Theory and Practice. Faruqui. Roozbehani. and S. eds.org/news. 38 North American Electric Reliability Corporation. see A. GA. Fox-Penner. and Sergici. Holland and E. 2010). MA.” Energy Journal 27. CA: GTM Research.” The Energy Journal 28. submitted. Hledik. Mansur. “Price-Responsive Demand Management for a Smart Grid World. “Customer Risk It Comes to Demand Response. see note 6 see note 2 above.” 46 above. “The Short. MI. Hledik. J. 172 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . 2010 U. 2010). Dahleh. Faruqui. “Volatility of Power Grids 2010 (Norcross. S. and J. and S. 3 (2005). The Costs Freeman. B. no. 32 H. Is FERC Its Own from Real-Time Retail Electricity Pricing: Bill Worst Enemy?” The Electricity Journal 22. Comments on PJM Demand Response Proposals. 37 Federal Energy Regulatory Commission. 82–109. and Electric Power Research Institute. Cheng. see note 39 above. Borenstein. 31 Demand Response Compensation in Organized Wholesale Energy Markets. Neichin and D. 2009 Load Impact 51 and Benefits of Smart Meters for Residential Evaluation for Pacific Gas & Electric Company’s Customers. L. 2 (2007): 111–130. and M. W. 8 Volatility and Hedgability. DC: Edison (2006).” Borenstein.” IEEE Transactions on Power Systems. and S. 93–116. Faruqui.S. see note 47 above.” presented at 49th IEEE Resource for Utilities and Grid Operators Conference on Decision and Control. August 24.brattle. Seventeen Pricing Experiments (San Francisco. Mitter. GA. Form 10-K Annual Report Fiscal Year Dahleh.” Quarterly 42 (2011). DC. Estimating A. Lansing. and Schwartz. 2010. Demand Response: A Multi-Purpose Real-Time Pricing. Kaldor.. Journal of Regulatory Economics 40. Hogan. smart-grid-newsletter/5599-surprising.

“An Experimental Comparison of Critical 70 Environmental Change Institute. A. Wolak. see note 9 above.. Faruqui. Donnelly. Sergici. Sharif. see note 40 above. J. “Backlash Against Smart Meters 77 A. Program.S. November 12. Literature Review for 69 Faruqui. Wood. and A. and Electric Power Research Energy Demand Research Project: Fifth Progress Institute. Smart Grid Consumer (Washington. Electricity Grid.” in Social Security Policy in a Changing Electricity Markets. and Laitner. note 36 above. Office of the Gas and States. and S. Ibid. 71 Borenstein. 62 75 For example. 2010. and J. 79 Faruqui. 74 Association of Home Appliance Manufacturers. The Impact 67 of Dynamic Pricing on Low-income Customers. Residential. see note 8 above. 2009). Demand%20Research%20Project%20Final%20 2009). 60 (Washington. see 54 note 10 above. note 40 above. “Implementing Opt-in. Industry’s Principles and Requirements for Achieving a Widely Accepted Smart Grid” AECOM. see note 45 above. February 13–16. 61 Manufacturers. Dynamic Electricity Pricing: Insights 57 M. IL. York: Environmental Defense Fund. Faruqui et al. 2010). IL: University of Chicago Press. see J. tion: A Survey of the Experimental Evidence.” Journal of Economic Perspectives. 65 K. Beshears et al. “‘Smart’ Meters Draw Complaints Winter Meetings. Brown. Palmer. and Laitner. of Informational Feedback on Energy Consump. 59 Ehrhardt-Martinez. Faruqui. the Energy Demand Research Project (Oxford. forthcoming. Analysis. Faruqui and L. Faruqui. 2011). white paper (Washington. S.K. 73 Allcott. and Leeds. DC: Association of Home Appliance U. and Association of Regulatory Utility Commissioners T. S.pdf. 58 Allcott. Behavior and Energy Savings: Evidence from Economics and Psychology” (PhD disserta- from a Series of Experimental Interventions (New tion. The Costs and Benefits of Smart 64 Smart Grid Consumer Collaborative. Ehrhardt-Martinez. and R. DC: Edison Foundation Institute Collaborative (2011). 2010). 78 P.” Energy 35 (2010): 1598–1608. see note 56 above. Darby. Wood. 2008). Liebman. DC: Edison Foundation Institute for Electric Efficiency.. eds. see note 67 above. and Schwartz. and Palmer. 2007). of Inaccuracy. Hobbs. white paper in Consumer Engagement.” Benefits of Smart Meters. Environment. 2010).K. Excellence Meters for Residential Customers. see note “Smart Grid White Paper: The Home Appliance 10 above. Faruqui. Management Systems for Demand Response Applications (Palo Alto.” presentation at National Santa Cruz Sentinel. see note 40 above. J. Bushnell. 2010. “IEE Releases: The Doesn’t Ease with New. Joskow. and Schwartz. Sergici. Sergici.gov. September 3. Donnelly. see note 39 above. 76 A. Chapter 7: Engaging Electricity Demand 173 . University of Peak and Hourly Pricing: The PowerCentsDC Oxford. Wood. http://www. “The Importance of Default Options for Retirement AECOM. and uk/Sustainability/EDRP/Documents1/Energy%20 D. Wise (Chicago. and Schwartz. Innovation in Retail Electricity Markets: The Overlooked Benefit (Chicago. see sgccs-excellence-in-consumer-engagement-study.” The New York Times. Davis. Office of the Gas and Electricity Markets. Berkeley. 53 68 Borenstein. CA. Alexander. 2009). 66 National Economic Research Associates.ofgem. J.” mimeo (Palo Alto. Assessment of Residential Energy Report (London. page 4. “The Impact 55 University. UK: F. and Wolak. Independent Reassurances. 2011. see note 56 above. July 2011). Zeller Jr. http://smartgridcc. 2010). Wood. University of California. AECOM. see 63 For example.org/ for Electric Efficiency. see note 33 above. 2011). “Creating a Smarter U. Energy Demand Research Project: Final 56 Savings Outcomes: Evidence from the United Analysis (London: U. 72 For example. CA: Stanford A. Letzler.

174 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .

The chapter starts with an introduction to the primary objectives and current practices that guide setting electric utility rates. Next.4. Finally.3 assesses tools that might improve regulation of the U. the regulatory investment easily could double if utilities deploy systems and policies that oversee the U. Section 8. We first find that consistent cost and performance metrics would allow for better comparisons of utility performance over time. and accommodate new types of significant amounts of capital over the coming generation. Recovering a substantial portion of distribution and transmission network costs through volumetric kilowatt-hour-based rates distorts both utility and customer incentives and over time may create implicit cross-subsidies across subsets of customers.1 gives an introduction to some of the most important distinctions and provides background for the findings and recommendations that follow.1 To decades to replace aging assets and expand the deliver on these promises in the most efficient network to meet incremental load growth. and we encourage continued experimentation and regulatory innovation in this direction. to improve system operation. beginning with the changing nature of utility investments. the impending changes will escalate regulatory policy design to higher prominence.2 also discusses the stress that rising rates can put on regulatory processes.S. and demand response. Although all forms of utility oversight generally reflect a common set of objectives. Section 8. The U. and we recommend that fixed transmission and distribution network charges be recovered largely through customer-level fixed charges. grid new transmission and distribution technologies also must be modernized. Many of the grid investments described in this report present greater risk than has been common over the past several decades. Issues specific to transmission were discussed in Chapter 4. the details vary for different types of utilities. we discuss customer incentives in the context of traditional rate structures. electric power industry must invest quality. and recommend that regulators collaborate to develop and publish such metrics. with particular attention on the determination of retail rates and treatment of investments.2 discusses emerging challenges for regulatory policy. Section 8.  While various reforms of public utility regulation have long been debated. load. we find that alternatives to volumetric charges are needed to mitigate distortions common to most utility rate structures.Chapter 8: Utility Regulation The challenges and opportunities described in earlier chapters place increasing burdens on electricity system regulators. Second. That and cost-effective manner. enhance service Chapter 8: Utility Regulation 175 . Section 8. and the potential tension between a regulated franchise monopoly mind-set and distribution system innovation. the discussion in this chapter focuses on utility regulation in general terms. and that those with utility oversight authority ultimately tie utility outcomes to their performance on those metrics. These inform the conclusions and recommendations summarized in Section 8.S. we find that new mechanisms for risk allocation and compensation are needed to balance incentives for risky investment with investment cost discipline. electric power system in the context of the challenges outlined in the first two sections.S.

S. of electricity systems in the U. electricity the cost of generation or wholesale power systems.” City councils or indepen- and some states enabled competitive retail dent boards oversee government-owned service. and various other sorts of (IOUs) that provided generation. we will refer to all AND PROCESSES such entities as “distribution utilities. utilities. particularly for larger commercial and utilities. The principles that energy that is managed by the ISO. varies widely. Generators in ISO regions governed through committees or boards sell into competitive wholesale markets for comprised of their members. Regulators universally determine the organizational. delivery of electricity from the high-voltage s Operational Efficiency: Utilities should transmission network to end-use customers deliver the quantity of electricity that generally is assigned as an exclusive franchise to consumers wish to purchase at the lowest a government-owned or cooperative enterprise reasonable cost while providing acceptable or to an IOU subject to some form of price reliability and other aspects of performance. In this chapter. First. the prices for franchised monopoly operators (ISOs) were established in much of services generally are determined administra- the nation. generation and transmission assets—including Until the 1990s. public utility commissions regulated. but most forms. distribution services are universally Third. limit generalizations across U. A number of states in ISO regions tively. Depending on the jurisdiction.S. supervising publicly and cooperatively owned The considerable heterogeneity in ownership. government and cooperatively owned systems are smaller utilities that purchase most or all govern rate determination are broadly similar their energy and transmission services from across each of these organizational forms. the industry was dominated by transmission lines. Even in areas a common set of objectives: that have restructured their electricity markets. power producers. but some commonalities exist in core purchases and transmission is passed through aspects of price determination. They provide only distribu. the hardware and software vertically integrated. investor-owned utilities in control centers. known as “rate cases. Customer-owned cooperatives are industrial customers. and distri- bution utilities and retailers purchase energy The principles that govern rate from those markets. State public utility commissions regulate restructured their electric utilities by separating the prices IOUs charge through proceedings generation from transmission and distribution. determination are broadly similar Some of the larger government-owned systems across each of these organizational were and are vertically integrated. 176 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .” recog- nizing they may be parts of vertically integrated As described in Chapter 1. equipment—also may be owned by entities of and distribution services under rates that state all these types. the organization organizations. according to rules that vary across regimes. utility oversight generally reflects treated as a natural monopoly. In the late 1990s. and regulatory forms allowed cost recovery for distribution services. vertically integrated firms or independent When we refer to “regulators” in this chapter. we generally include entities responsible for tion services to customers in their jurisdictions. independent transmission system Second. 8. market. transmission.1 REGULATORY OBJECTIVES regulation.

and other non-IOU systems typically determine rates through different processes.3 incentives for cost minimization. minimally for any given utility or at any point in time. s DETERMINING THE UTILITYS TOTAL COSTˆINCLUDING a fair rate of return on prudent capital One of the greatest challenges facing regulators investments—which forms the basis of a total attempting to implement these objectives is “revenue requirement”.2 i Municipal.s Dynamic Efficiency: Utilities should make Two extreme theoretical regimes illustrate this efficient investments in innovation so that problem. utility costs. and asymmetric information: while a utility’s demand characteristics and opportunities for s SETTING RATES BY ALLOCATING A SHARE OF REQUIRED cost reduction and investment may be funda. retail electricity energy) or other regulated costs. distributed generation. These schemes s Consumption Efficiency: Customers should ensure full cost recovery bear the incremental cost that their decisions because the regulator can One of the greatest challenges impose and be given appropriate incentives generally measure actual facing regulators attempting to consume electricity only when its value to costs precisely.S. they may greatly exceed costs or be far below them should do so in a cost-effective. or electric vehicles.4 Chapter 8: Utility Regulation 177 . distortive manner. This provides strong incentives for cost reduction because a dollar of cost savings s Other Policy Objectives: Where utilities are translates into a dollar of profit. But they them is at least as great as the incremental provide no incentives to to implement these objectives cost of producing and delivering it. high-powered. this makes it impossible to simultaneously ensure These processes play an important role in exact cost recovery and provide optimal determining utility and customer incentives. for compensation for public capital for instance. In a textbook price cap regime. with system cost recovery and investment In contrast. As a general matter.i The mechanics of this regulatory process may include enhanced energy efficiency and divide into two broad tasks: accommodation of renewable generation. Regulators frequently establish distributional goals. with “cost” understood to include regimes decouple a utility’s revenue from its a fair rate of return on capital to investors or actual costs. should be set at the lowest level consistent to minimize those costs. prices are set once and remain fixed non-IOUs. but they lead to similar outcomes as cost-of-service regulation and involve similar efficiency problems. customers who are thought to be better able processes based broadly on cost-of-service to pay bear a larger share of network (non. Prices managers to exert effort is asymmetric information. Pure cost-of-service regimes set utility they are able to meet future demands at the revenues at all times exactly equal to observed lowest reasonable cost. But revenues expected to support other policy goals. revenue to each customer class and deter- mentally uncertain. incentive-based incentives. does the regulator. such as While actual regulatory practices depart from requiring firms to design rates such that these theoretical extremes in significant ways. including a normal return on investment in the rate base. over time. cooperative. Other goals rates. regulation determine most U. managers of the utility mining the structure of rates to recover that typically have better information on these than revenue.

typically are first allocated to customer classes because of different views of the optimal level based on such factors as connection voltage and of reliability—most traditional distribution the class’s estimated contribution to the local system investments have been relatively low peak demand. a portion of distribution service investments by IOUs. total cost allocated to each customer class to Generation and transmission projects typically expected revenue. This oversight kWh energy consumption. 178 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . in energy costs generally flow through to rates regulators of various sorts must give their via frequently changed volumetric energy approval before entities can build new trans. charges applied to each kWh mission lines. the retailer determines the contractual energy price. the regulator must investments required to provide acceptable decide how to set rates to recover those costs. In investments. For large customers. receive detailed project-specific reviews.ii In some regulate the returns to rate-based transmission jurisdictions. distribution and transmission system costs are recovered from small commercial and ii In competitive retail markets. As discussed in Chapter 4. prices. during some prior base period or its contracted demand for some future period. based on historic consump- involve more significant outlays and. One important way the practice of cost-of. tion levels and usage patterns for that class. Similar reviews apply electricity consumption and change with fuel to capital investments by publicly and coopera. and reviews generally have been routine. charges—that is. Rates are computed to match the risk. We refer to network costs—the capital and service regulation differs from the theoretical operating costs of the distribution and trans- extreme characterization is that only the costs mission network—as “fixed costs” because they that regulators deem prudent are recoverable are largely unaffected by short-run changes in through the ratemaking process. All customers face distribution rates that are at least partially volumetric. ingly. regulators Once the required revenue for cost reimburse- develop familiarity with the types and levels of ment has been determined. Fluctuations tively owned utilities. These costs are also process is particularly important for capital relatively stable between rate reviews. While utilities and regulators tion. Generation investments by IOUs contrast. and state commissions and the consumed—often implemented by automatic Federal Energy Regulatory Commission adjustment formulas indexed to costs. which is combined with the regulated transmission and distribution charges into a single customer bill. costs of the distribution and transmission network— and most of the allocated costs are recovered through fixed monthly charges. among other factors. rates also may be subject to an automatic adjustment between rate cases. typically based as “fixed costs” because they are largely unaffected on that customer’s actual peak demand level by short-run changes in kWh energy consumption. Experiential learning is important for the prudency review process: over time. network costs at the distribution level sometimes may disagree—for example. energy costs—reflecting either fuel outside competitive wholesale market areas costs for vertically integrated utilities or and distribution system investments for all wholesale power acquisition costs for distribu- IOUs are subject to state commission review tion utilities in restructured markets—vary with and rate regulation. We refer to network costs—the capital and operating volumetric distribution charges tend to be low. In contrast. accord. service for a given utility system and set of To determine customer charges for distribu- technologies.

the delivered electricity rates are lead to cost-effective decisions.7 But changes in the Revenue adequacy for the distribution utility economic environment and the demands is met by matching the expected aggregate placed on utilities are likely to increasingly revenues across all customers.” with demands on electricity networks. consumption rises above threshold levels (a is likely to require electricity rate increases.residential customers almost entirely through incentive-based regulation. significant investments in new technologies Many customers receive bills that aggregate that systems may need to make to improve energy. with very inflation environment. Paying for constant within a billing cycle for most residen.2 GROWING CHALLENGES rates generally are then fixed until the next FOR REGULATORY POLICY regulatory review. the annual average both by limiting demand engagement. as increase in the Consumer Price Index was only discussed in Chapter 7. to the revenue the coming decades. transmission. each of which may approaches have difficulty assuring that incen- change across usage “steps. Chapter 8: Utility Regulation 179 .5 A variety of of their share of network costs onto other factors contributed to this performance. in networks. and incentives markets and. likely to face rates that vary with time of day is likely to require electricity rate increases. and average retail electricity rates increased customers the incentive and ability to shift part by only 2. such as renewable portfolio standards. Larger higher power costs that may arise from policy commercial and industrial customers are more mandates. Finally. In some jurisdictions. Several factors will combine to increase the likely stress on existing regulatory practices in The detail of charges customers see on their the coming decades. In another departure from the theoretical model of pure cost-of-service regulation. innova- them efficiency gains in generation as a conse. skill set. in previous chapters. apart from the impact of any automatic adjustment formulas. Many will arise from the bills varies tremendously across jurisdictions. such as renewable portfolio standards. and for some. often with two or more steps that increase the kWh charge as monthly mandates. Traditional regulatory mission. notably those with retail system operation and accommodate new competition. and distribution.” As noted in tives and compensation for such investments Chapter 7. bills may be “unbundled. increased network investment. increasing the value of required to match total allowed system costs. trans. Between 1990 and 2010. combined charge and usage charges based on kWh with higher power costs that may arise from policy consumption. Prevailing rate structures may compound the difficulty of managing increases.1% per year on average. as described separate charges detailed for energy. as we describe below. among customers. systems. unlike some earlier periods when costs were rising more quickly. tion in customer-facing activities may not be quence of competitive wholesale electricity well-matched to the focus. those 8. and distribution charges into a (generally small) fixed monthly customer Paying for increased network investment. with system Rising rate levels challenge most regulatory conditions.6 In this low- volumetric delivery rates ($/kWh).6%. given rates and strain the existing regulatory system during expected levels of consumption. all jurisdictions. because of of traditional utilities or their regulators. utility ratemaking low fixed monthly customer charges in virtually overall has functioned relatively smoothly. and day of the week. and by giving some 2. combined with tial and small commercial customers. regulatory adaptation and innovation. structure known as “increasing block pricing”).

Complicating calculations further. As a result. uncertainty through ex ante prudency reviews. and and 5). utilities would invest in risky new achievability. these technologies if the expected returns when the investments may require coordination across technology is successful are at least sufficient to different utility business units and the integra- compensate for possible losses if the technology tion of legacy data communications and information management systems. 4. and 6). may be subject to uncer- and priced. modernization invest- might be eager to undertake investments of all ments may not be easily justified by predictable sorts. This suggests that regulated utilities decisions. however. distributed nology. and electric vehicles (Chapters 3. efficient investments. investments. The total cost of integrating large system benefits of emerging technologies can amounts of variable energy resources and be difficult in such situations.8 Electric vehicles may require distribution For example. the uncertainty associated may have limited experience with these tech- nologies and have to work closely with equip- with new technologies and risk aversion on the part ment vendors that may have little experience of regulators and utilities may discourage even with electric power distribution systems. sion and distribution systems in a cost-effective manner will require utilities to evaluate and Utilities considering significant capital invest- potentially adopt technologies that significantly ments generally seek to reduce the regulatory depart from traditional grid investments. many new technologies have benefit streams that poten- fails. regulators may have a hard time deciding on the degree of automation and active whether to approve investment proposals and management of the transmission and distribu. deployments in existence to serve as models. net change in operating costs. infrastructure (AMI). cost-of-service regulation tially will extend many years after costs have encourages utility capital investment by been incurred and are partially a function of assuring utilities a “fair” rate of return on their future technology innovation and deployment investment. ditures. such as those necessary to effi- timing and magnitude depend critically on the ciently and reliably integrate distributed penetration and geographic distribution of generation or effectively use the wealth of electric vehicles within the distribution system information provided by advanced metering and how charging is metered. tainty about not only the level of costs and benefits but also about their timing and Ideally. The Nature of Investments Is Changing and risk aversion on the part of regulators and utilities may discourage even efficient Meeting the future expectations for transmis. and there are no system-scale estimates of costs and benefits. In theory.9 Forecasting the level of capital expen- generation. however. experience assessing the tech- of variable energy resources. investments in new distribution network component upgrades. Among the required new capabilities are But for many new transmission and distribu- automated sensing and control (Chapters 2 tion technologies considered in this report. if any. the short-term improvements in reliability or uncertainty associated with new technologies incremental improvements in operations or 180 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . development of more responsive neither the utility nor the regulator may have demand (Chapter 7). and accommodation much. To be most effective. Utilities In practice. In practice. how to hold utilities accountable for their best tion networks. risky or not. although their technologies. Without good distributed generation in particular will depend data. controlled.

Extrapolation or meta-analyses conservatism may dramatically retard the of pilot results may help to narrow the range of adoption of technologies to modernize the grid. may increasingly impede efficiency may create similar conservatism among regula. residential and small commercial customers vary able criticism for approving rate increases for technologies that fail to meet expectations or. some as may occur during peak demand periods for of the benefits may accrue to customers in the customers facing time-invariant rates based on same region who are not served by that distri.12 Such Chapter 8: Utility Regulation 181 . even among larger customers. involve well-publicized problems. For example. are rare. Rates charged to most about. average costs.11 tion. In contrast. because the costs were to modernize the grid. even among larger customers. if a particular distribution utility costs are rare.operating costs. Regulators are unlikely to face censure for in Chapter 7. dramatically slow cost-effective investments Colorado. are familiar to utilities and regulators. Real-time prices that reflect current system costs worse. objectives over the coming decades. likely outcomes. Traditional Rate Structures Distort Incentives aversion to such risks may make utilities excessively cautious in proposing investments The traditional structure of electricity rates with highly uncertain costs. The political process in the U. state regulators have sometimes the system. substantially in excess of those initially antici- pated.S. technologies and systems may end up incurring This can increase energy acquisition costs and costs that are avoided in later adoptions through require additional investments in both genera- learning spillovers from early experiences. after the regulator had approved the investment proposals. but considerable uncertainty even when deployment of those technologies for system deployments may remain. a recent Colorado Excessively conservative decision-making decision denied Xcel Energy cost recovery for by utilities and their regulators may certain smart grid investments in Boulder. increases the responsiveness of its customers’ demand or the efficiency of their energy use and If delivered prices are below incremental costs. thereby reduces regional generation costs. customers will use inefficiently bution utility. leading of new investments is that some significant to excessive consumption during system peaks benefits of grid modernization may not be fully and inefficiently discouraging off-peak consump- internalized by a single utility or its customers. many consumers face rates that failing to encourage novel investments that often deviate substantially from the incremental might have yielded new benefits or moderated costs their consumption decisions impose on rate increases but that customers never learn the electric power system. Another challenge posed by the changing nature little or not at all over the hour or the day. is likely to be the most cost-effective path to accommodate the policy goals impressed on In addition. approved investments in ex ante prudency reviews only to later deny a utility’s full cost FINDING recovery.10 While such actions may be needed to provide strong incentives for managing costs. Similarly. regulators may face consider. early adopters of new high quantities of electricity at high-cost times. Real-time prices that reflect current system For instance. As discussed tors. reducing capacity utilization and increasing These considerations may bias investment average costs (and therefore average rates) toward the mature technologies and assets that above efficient levels. tion and grid capacity to meet peak demand.

restructuring and retail competition so they iii Some have suggested that distortions induced by volumetric charges to recover fixed costs and the potentially greater distortions caused by increasing block pricing may be justified by the presence of imperfections in the market for electricity—for example. Volumetric charges generation may receive additional direct are especially problematic when energy subsidies. proposed “zero net energy” subsidy to self-generation of all sorts. with a healthy skepticism toward the benefits of adding layers of distortions. or diesel. particularly from “dirty” sources. High per kWh distribution rates create an implicit for example. rate practices give utilities customers. But prices that are higher than incremental could shift their purchases to lower-cost costs can be equally distortionary. Such decisions must be based on careful analysis and situation-specific modeling and measurement to ensure improvements to welfare. Consider.15 While the economic theory of “second-best pricing” recognizes that marginal cost prices may not be optimal in the presence of market and decision-making imperfections. because cost-based prices do not reflect the incremental social cost of electricity consumption. If network costs are largely use of self-generation. buildings: if network costs continue to be recovered on a per-kWh basis.” Forty-six states would expect.13 The ering a substantial portion of distribution and coming decades could see a similar phenom- transmission costs through volumetric charges. Inclining block pricing used in some jurisdictions—rates On the customer side. there is no theorem suggesting that one potential distortion merits another. reduced consumption decreases and the District of Columbia use net-metering revenue by much more than it decreases costs. While these problems are not new.iii The distortions incremental cost of providing that energy. enon among smaller commercial and residential the dominant U. when the average that increase as monthly consumption rises— delivered price of electricity is higher than the can exacerbate the problem. This could lead to inefficiently high consumption falls. while paying little or and commercial customers in states with nothing toward the cost of the system or the high regulated rates in the 1990s to press for option to use the network. High per kWh distribution rates incentives to increase their sales and discourage create an implicit subsidy to self-generation of energy conservation and distributed generation all sorts—solar. wind.16 182 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . drawing and injecting generation costs led many large industrial power on demand. they customer are not likely to fall with their use of are likely to grow over the coming years as distributed generation.S. This practice in effect public policies increasingly favor energy adds an extra subsidy to distributed generation efficiency and distributed generation. wholesale providers or self-generate. Renewable because they reduce sales. and there is no reason to unchanged by consumption changes. these customers purchases of power from the regulated utility. systems that compensate distributed generation creating a revenue adequacy problem for the through avoidance of volumetric retail rates14 utility and perhaps setting off a spiral of rising even though network-related costs to serve a rates. or because consumers need higher prices to overcome their inertia and make desirable efficiency investments. as we assume it will all be “clean. not given to grid-scale generators. could in theory receive all the benefits of being This situation with respect to wholesale connected to the grid. there caused by these implicit subsidies rise with the is an incentive for “disintermediation”—reducing penetration of distributed generation and with energy conservation more generally. By recov.

may lead tion or efficiency investments. mission and distribution investments. as even increase them as investments are needed noted in academic work in the mid-1970s and to accommodate more than modest increases reinforced by the recent backlash against higher in renewable generation. The Electric Power Research Institute not. This raises questions of both horizontal (EPRI) estimates the average incremental increase inequity—treating otherwise similar consumers in monthly electricity bills required to finance differently—and vertical inequity—penalizing smart grid transmission and distribution lower-income consumers. customers’ total bills after rates adjust will be when nominal rate increases were modest and smaller only by the energy cost differential. would seem maintaining the smart grid system may require difficult to rationalize on equity or political higher ongoing investment levels than in recent grounds. and significant problem over the past two decades. programs do not appreciably reduce transmis- Utility customers. Who pays for the monthly bills coincident with AMI rollouts in revenue lost because of such programs under Texas and California.18 This was not a consumption. Requiring middle.19 In the consumption. distributed generation and efficiency Rising costs often test regulatory systems. But the increased cost of new trans- investing in self-generation or efficiency. Rising Rates Stress the Regulatory Process Moreover. so as traditional rate structures do. If only some customers reduce as renewable portfolio standards or more their net energy purchases through self-genera- stringent environmental regulations. Chapter 8: Utility Regulation 183 .17 Customer resistance to current rate structures? If all customers reduce higher rates may pressure regulators to avoid consumption proportionally and fixed system increases.FINDING Requiring middle. such dissonance. distribution to significant increases in the delivered cost of system costs will be shifted onto those who do electricity. particularly at the residential sion or distribution system costs—and may level. If average real prices fell across most electricity consumers do not understand this when markets. do historic distribution system assets. who may be dispro- investments will be in the range of 8%–12% portionately represented among those less able for residential and commercial customers to finance investments to reduce net electricity when amortized over a 10-year period. even when they are needed to costs are unchanged. These expenditures are in addition to rate increases needed to compensate for higher generation costs and investment in transmis- sion and distribution to meet load growth. as traditional rate structures do. would generation that is not given to grid-scale seem difficult to rationalize.or lower- same report. generators. EPRI also notes that digital-based income customers to subsidize wealthier technologies depreciate more quickly than households’ investments in energy reduction. are sensitive to higher nominal bills. in the long run volumetric compensate utilities for efficient levels of rates must rise to exactly offset the fall in operating and investment costs. decades. consistently smaller returns than anticipated combined with possible increases in average from their investments may create substantial generation costs due to policy mandates.or lower-income customers to Net metering policies provide an implicit subsidize wealthier households’ investments in energy subsidy to all forms of distributed reduction. The per kWh rate will rise.

will be critical to functions provided by distribution system offsetting some of this pressure. their cost of capital may rise. This is essentially the approach used in dissatisfaction and may distort consumption Texas. regula. particularly if destabilization: If utilities make investments in consumers focus on nominal rates. necessitates rate crisis of 2000–2001. exacer. Finding ways to distribution networks comprise a natural increase the efficiency of the power system. while more than half of U. Indeed. replace end-of-life assets. a number of countries and in many U.S. states Communicating the rate implications of new was the perceived potential for competition in investments and programs poses challenges. small commercial customers may further reinforce a perverse dynamic over time. to a legal monopolist. As Chapter 1 noted. both wholesale generation and retailing however. customer-side investments on the basis of illusory bill savings may create considerable Without policy adaptation. This. A “Franchise Monopoly” Mind-set May bating cost-recovery challenges. Higher prices offset the savings tition also varies widely. or transparency and active customer education at all. While there is broad consensus that physical tors. Without retail 184 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID .22 The form of retail compe- increases. and policy makers. states allow retail competition for large The volumetric rate structure for residential industrial or commercial customers. from what they would otherwise been but still involve rate increases may create have been. For example. they may be unable to meet the conflicting demands of customers.S. implemented efficiency or conservation decisions and may through its Electricity Directive. there is little agreement that other and therefore reduce costs. in turn. While some have argued that full unbundling constitutes the exception and increasing efficiency and distributed generation not the rule in the U. monopoly. a primary and engagement also may help mitigate some motivation for electricity restructuring across customer resistance to necessary rate increases.S. inducing renewables. the system risks political backlash over time. few today and small commercial customers may further have competitive electricity retail markets at reinforce a perverse dynamic over time. substantial dissonance if customers anticipate consumers receive regulated delivery services rate decreases that fail to materialize. Improved operators should be assigned exclusively. but retail competition associated with decisions. the residential level. But if utilities Discourage Innovation defer investments. and ultimately elec- costs from what they would otherwise have tricity rates. from their distribution utility and are free to purchase energy and other related services The volumetric rate structure for residential and from competitive electricity retailers. in which the provide an even larger incentive for future distribution utility is allowed to sell only reductions in customer demand. investments that reduce markets to reduce costs. energy and other prices up still further. ratcheting distribution network services. At the extreme is the customers may have expected from their fully separated EU model.21 Under retail competition. In part. this reflects the Reduced consumption threatens the ability of a retrenchment from electricity restructuring distribution utility to recover its predominantly efforts that followed the California electricity fixed costs.20 an environment of substantial uncertain cost recovery. This cycle of rising services must be purchased from other enter- distribution rates further incites customer prises. and integrate are benefits of volumetric rates.

Innovation responses. Inherent in their design is the possi- over who owns and pays for customer premises bility that utility revenues may exceed or fall demand management technologies and short of costs for significant periods of time. electricity inclusion in a utility’s rate base. natural gas distribution.S. leverage retailing regulation of traditional “public utility” efficiencies by facilitating operations across sectors—for example. local wireline telecom- many distribution systems. innovation in retail electricity more widespread use of new regulatory tools. This section may impede the efficient development and discusses a set of the most promising regulatory introduction of new services. But rate regulation and continued experimentation. stimulate the development of new Incentive Regulation energy management systems. utilities to reduce operational costs without sacrificing reliability or quality of service. When energy efficiency and savings becomes a regulatory priority. from cost-of-service mainly to customers..competition. In the presence of new and emerging technolo- gies. reflecting 8. Chapter 8: Utility Regulation 185 . whether the cost of those investments should The potential savings from these approaches be “socialized” across all customers through may be considerable. restructured U. services depends on incumbent monopoly significant adaptation of regulatory processes. regulated firms have little to gain from voluntarily Regulators around the world have shifted regulation of assuming risks: if things go well. by shifting distribution companies in England and Wales those responsibilities onto competing retailers. the to incentive-based schemes designed to motivate pain is borne mainly by managers and shareholders. for Enhance Performance-Based example. provide invest- ment funds for deployment of new customer Regulators around the world have shifted premises technologies.iv that were subject to price cap regulation dramatically reduced operating costs. vate utilities to reduce operational costs ment activities.3 HOW SHOULD POLICY RESPOND? the combined benefit of incentive regulation and privatization. that the main business of retailers is to sell electricity. Addressing these challenges generators instead of cost-of-service regulated in a cost-effective manner is likely to require rates. electricity markets reveals gies. distribution companies. and policy emphasis on energy ated with some forms of regulatory incentives efficiency and alternatives to conventional and with market-based competition among generation sources. however.24 iv Regulators cannot ignore.. and provide a better munications. if things go badly. For example. and match for the expertise and talent required to distribution of electricity—from cost-of-service be successful in services that rely heavily on to incentive-based schemes designed to moti- information technology and customer engage. the benefits go traditional “public utility” sectors . are best suited for this job. Separating retailing from without sacrificing reliability or quality of distribution services also could resolve debates service. penetration of distributed generation and improvements in generator efficiency associ- electric vehicles. energy service companies. involves risk. and as we have noted. competitive entry into retailing may.S.23 While the use of incentive- Challenges to regulatory policy are gaining based regulation in U. empirical analysis of of innovation in distribution system technolo. whose only purpose is to reduce consumption or to shift it to low cost periods of time. electric utilities has increased prominence with the quickening pace been more limited.

Assessing the impact of – X regulation. where RPI refers to the Retail performance-based regulation across U.S. plus or regulation that included explicit penalties or minus a fixed percentage adjustment factor set rewards based on performance in reliability by the regulator. incentives. in Index). revenue cap for delivery service that is fixed over a given time period. In 2001. and performance reporting. An additional 23 states set service period considered. such as integration of a number have returned to more traditional distributed generation and electric vehicles cost-of-service-based ratemaking supplemented or demand response. Consumer Price distribution utilities is extremely difficult. Candidates for measurement and a broad range of changes. regulators in many tives for efficient operations. power at least 28 distribution utilities in 16 states were quality measurement. and incentives. outages.27 other countries in Europe and Latin America have evolved to include performance metrics Jurisdictions that require electric utilities to and assign penalties or rewards depending on report performance on various metrics have whether utilities meet these predetermined commonly used the results to provide incen- targets.26 Of these. taking into account expected targets or required utilities to report perfor- productivity gains. firms are allowed to charge systems involving quality of service. part because of this heterogeneity in standards.1 classifies U. two states generally set to equate the discounted expected adjusted the allowed rate of return based on values of costs and revenues over the time performance.S. from relatively reporting would include detailed information short-term price caps. suggesting reduce spending on reliability or service quality. Expanding their use and formally regulation that included explicit penalties or rewards tying performance to financial incentives can based on performance in reliability and/or quality increase those gains. prices (or earn total revenues) that increase 16 states had a form of performance-based with an index of the general price level. customer service indicators. Internationally. states began experimenting with a variety powered than the price or revenue caps many of incentive-based regulations. and utilities are faced with assessment of new technologies. During the 1990s. 16 states had a form of performance-based level currently employed across many U. 186 MIT STUDY ON THE FUTURE OF THE ELECTRIC GRID . The adjustment factor is and/or quality of service. and X is the adjustment factor deter. rate case on the duration. typically three to five Figure 8. incidence. pioneered in mance but associated no financial consequences England and Wales. These included European systems employ. and earnings sharing plans. units incentives to meet certain heat rate and availability goals are associated with improve- Since one way of reducing costs may be to ments in generating unit efficiency. mined by the regulator. utilities. Under this system. at least some role for performance-based implementations in England. While less high- U. This approach.S. and cause of moratoria. generally is known as RPI to those reports.25 Subsequently possible policy goals. and is thus independent of actual costs.S. particularly as regulators of service. One regulated firms can increase profits by reducing study finds programs that provide generating costs relative to the index. adoption of performance-based regulatory During this period. Price Index (like the U. As of 2005. states based on their years. the most common form of by reliability or quality-of-service targets with incentive-based regulation sets a price or total penalties and rewards as performance incentives. to rate freezes. Australia. performance-based incentives may yield some gains even at the As of 2005. and performance on regulated under such mechanisms.S.

29 regulatory jurisdictions. State of Distribution Reliability Regulation in the United States (Washington. can improve the efficiency of the distribution system and the quality of Regulators in the European Union and else- where use two tools—benchmarking and utility investments. or target only (neither penalty nor reward) Reporting Only – utility is required to report performance with no penalty.28 Regulators outside the U. commonly benchmark operating costs and The value of cost and performance data could quality of service across different types of be substantially enhanced by coordinating the networks to determine efficient cost and definition and measurement of metrics across performance levels for comparable utilities. data from a large number of potentially similar utilities. Number of States in Each Category reward or target Source: Davies Consulting. this regulation. 2005). While heterogeneity across significant cost drivers external to the utility service territories may preclude simple formu.  Better and more reference network models (RNMs)—that can comparable data on system costs and further enhance the value of incentive-based performance. commonly adjustment factor that satisfies expected publicized metrics could support incentives for cost-recovery goals. Inc. improvement and facilitate the development of the most cost-efficient utilities do not always Chapter 8: Utility Regulation 187 . FINDING standards of service. DC: Edison Electric Institute.1 States with Performance-Based Cost-of-Service Regulation. Progress along these Greater performance-based regulation directions should be strongly encouraged. Benchmarking involves gathering accommodates policy goals. 2005 No Requirements QoS Penalty QoS Penalty & Incentive QoS Target ROE-based PBR Reporting Only ROE-based PBR: 2 No Requirements: 12 QoS Penalty & Incentive: 3 QoS Penalty: 11 Definitions Return-on-equity-based Performance-based Regulation (ROE-based PBR) – rate of return is set with a dead band (range the utility and shareholders assume all benefits and cost) and a live band (range above and below the dead band that would have a sharing mechanism assigned) QoS Target: 11 Quality of Service (QoS) – rate of return is set by using the conventional cost of service methodology and the utility has reliability and/or quality of service Reporting Only: 12 targets set by the commission with either a penalty & reward incentive. and thus choose a price or revenue cap and laic comparisons across utilities. As might be expected..S. including how well the system regulation. penalty only.Figure 8. expanding their use in For prices set using RPI – X formulas. and publishing data on observed exercise can help regulators understand the performance. are needed for and analyzing comparable cost and network its effective implementation. Table 4-1.

S. once built and validated. distribution utilities tive insights into the impacts of potential currently make it difficult to use benchmarking system changes. can be shared across jurisdictions to provide specific data—particularly on performance— regulators with both qualitative and quantita- for a large number of U.32 But by incorporating lessons from networks. particularly for capital investments. RNMs thus can be design choices of underground or overhead used to assess capital and operating expenses lines. distribution utilities currently make realized operating and investment performance it difficult to use benchmarking methods. and they may improve it can be difficult to adjust for these and other our understanding of how these new network differences.33 By incorporating technical and reliability constraints. distribution network Hindering benchmarking efforts are differences configurations. data will need to be collected over a targeted experiments or pilot projects. regulatory judgment in assessing the impact ization proceeding at different rates across the of new policies and informing the design of nation. detailed. They also may be helpful in estimating Although econometric models combined with costs of nontraditional investments for inte- detailed system data may help regulators to grating distributed generation and accommo- expand the set of firms deemed “comparable.S. The model can complement methods. exhibit high levels of service quality. and accounting systems but also in the expansion of existing or depreciation practices across jurisdictions. represent evolving distribution systems. users affect reliability and quality of service. its network design. amount and not only in the design of new distribution timing of specific investments. RNMs permit simulation of investment customer base. benchmarking may at least range of technologies under consideration is reduce the significant uncertainties associated needed to build models that can accurately with investments in some new technologies. the more informative the results of be