Power Systems, Third Edition (part of the five-volume set, The Electric Power Engineering Handbook) covers all aspects o

*414*
*76*
*45MB*

*English*
*Pages xiv+542
[558]*
*Year 2012*

*Table of contents : Power Systems, Third Edition......Page 4Contents......Page 6Preface......Page 10Editor......Page 12Contributors......Page 14I Power System Analysis and Simulation......Page 161 The Per-Unit System......Page 181.1 Impact on Transformers......Page 211.2 Per-Unit Scaling Extended to Three-Phase Systems......Page 251.3 Per-Unit Scaling Extended to a General Three-Phase System......Page 292.1 Introduction......Page 322.2 Discussion of Per Unit Quantities......Page 342.4 “a” Operator......Page 372.5 Phase and Sequence Relationships......Page 382.7 Sample Three-Phase Fault Calculation: Fault at Bus T2H......Page 412.9 Negative Sequence Networks......Page 422.10 Sample Phase to Phase Fault Calculation: Fault at Bus T2L......Page 432.11 Zero Sequence Networks......Page 442.12 Sample Phase to Ground Fault Calculation: Fault at Bus T2L......Page 473.2 Power Flow Problem......Page 503.4 Formulation of Power Flow Equations......Page 523.5 P–V Buses......Page 553.7.1 Basic Power Flow Equations......Page 563.8.1 Newton–Raphson Method......Page 573.8.2 Fast Decoupled Power Flow Solution......Page 583.9 Component Power Flows......Page 59Further Information......Page 604 Fault Analysis in Power Systems......Page 624.1 Simplifications in the System Model......Page 634.2.1 The Balanced Three-Phase Fault......Page 654.2.2 The Single Phase-to-Ground Fault......Page 664.2.3 The Phase-to-Phase Fault......Page 674.3 An Example Fault Study......Page 684.3.1 Balanced Three-Phase Fault......Page 704.3.2 Single Phase-to-Ground Fault......Page 714.4 Further Considerations......Page 764.5 Summary......Page 77References......Page 78Further Information......Page 795.1 Power Flow......Page 805.1.1 Admittance Matrix......Page 815.1.2 Newton–Raphson Method......Page 845.2 Optimal Power Flow......Page 925.2.1 Steepest Descent Algorithm......Page 945.2.2 Limitations on Independent Variables......Page 1005.2.3 Limitations on Dependent Variables......Page 1015.3 State Estimation......Page 104II Power System Transients......Page 1126.1 Introduction......Page 1146.2.1 First Strokes......Page 1156.2.2 Subsequent Strokes......Page 1166.3.1 Ground Flash Density......Page 1176.3.2 Current Peak Value......Page 1186.3.3 Correlation between Current and Other Parameters of Lightning......Page 1196.4 Incidence of Lightning to Power Lines......Page 120References......Page 1217 Overvoltages Caused by Direct Lightning Strokes......Page 1247.1 Direct Strokes to Unshielded Lines......Page 1257.2 Direct Strokes to Shielded Lines......Page 1277.2.1 Shielding Design......Page 1307.3 Significant Parameters......Page 1327.4.1 Unshielded Lines......Page 1347.4.2 Shielded Lines......Page 1377.5 Effects of Induction for Direct Strokes......Page 1387.6 Ground Impedance and Corona under Lightning......Page 139References......Page 1408 Overvoltages Caused by Indirect Lightning Strokes......Page 1428.1 Inducing Voltage......Page 1448.2 Induced Voltage......Page 1458.3 Green’s Function......Page 1468.4.1 Evaluation of Green’s Function......Page 1478.4.2 Induced Voltage Caused by Return-Stroke Current of Arbitrary Waveshape......Page 1488.5 Induced Voltages on Multiconductor Lines......Page 1518.6 Effects of Shield Wires on Induced Voltages......Page 1538.7 Stochastic Characteristics of Lightning Strokes......Page 1548.8 Estimation of Outage Rates Caused by Nearby Lightning Strokes......Page 1558.A Appendix A: Voltage Induced by Linearly Rising and Falling Return-Stroke Current......Page 158References......Page 1599.1 Transmission Line Switching Operations......Page 1629.4 Shunt Reactor Applications......Page 166References......Page 16710 Very Fast Transients......Page 16810.1 Origin of VFT in GIS......Page 16910.2.1 Internal Transients......Page 17010.2.2.1 Transient Enclosure Voltages......Page 17110.2.2.3 Transient Electromagnetic Fields......Page 17310.3.1.1 Bus Ducts......Page 17410.3.1.4 Spark Dynamics......Page 17510.3.1.7 Three-Phase Models......Page 17610.3.2.1 Enclosures......Page 17710.3.2.2 Ground Straps......Page 17810.3.3 Testing and Simulation......Page 17910.3.5 Validation......Page 18110.4.3 Disconnectors and Breakers......Page 183References......Page 18411.1 General Concepts......Page 19011.2 Material Properties......Page 19111.3 Electrode Dimensions......Page 19211.5 Initial Transient Response from Capacitance......Page 19511.7 Ground Electrode Impedance: Wire over Imperfect Ground......Page 19711.8 Analytical Treatment of Complex Electrode Shapes......Page 19811.9 Numerical Treatment of Complex Electrode Shapes......Page 19911.10 Treatment of Multilayer Soil Effects......Page 20011.11 Layer of Finite Thickness over Insulator......Page 20111.12 Treatment of Soil Ionization......Page 20211.13 Design Process......Page 20311.A Appendix A: Relevant IEEE and IEC Standards in Lightning and Grounding......Page 204References......Page 20712.1 Fault Interruption Process......Page 21012.2 Analysis Principles......Page 21112.3 TRV for Transformer-Fed Faults......Page 21412.4 TRV for Capacitor Bank Switch Opening......Page 21512.5 TRV for Line-Fed Faults......Page 21912.6 TRV for Current-Limiting Reactor Faults......Page 22212.7 Switchgear Tests and Standards......Page 22312.8 TRV Mitigation......Page 226References......Page 22713.1 Arrester Types and Auxiliary Equipment......Page 23013.2 Ratings and Tests......Page 23213.3 Selection by TOV......Page 23513.4 Selection by Energy Rating......Page 23613.5 Arrester Modeling......Page 23813.6 Applications......Page 241References......Page 24514.1 Insulation Coordination......Page 24814.3 Probability of Flashover......Page 24914.4.1 Voltage Waveshape......Page 25014.4.4 Effect of Atmospheric Conditions on Air Insulation......Page 25214.4.6 Insulator Contamination......Page 25314.4.7 Application of Surge Arresters......Page 25514.4.8 Examples of Surge Arrester Application (Nonself-Restoring Insulation)......Page 256References......Page 258III Power System Planning (Reliability)......Page 26015.1 Introduction......Page 26215.2.1 Preparing for Competition......Page 26315.2.2 Present View of Overall Problem......Page 26415.2.3 Economic Evolution......Page 26615.2.5 Fully Evolved Marketplace......Page 26715.2.6 Computerized Auction Market Structure......Page 27015.2.7 Capacity Expansion Problem Definition......Page 27215.3 Regulated Environment......Page 27315.4 Other Sections on Planning......Page 275References......Page 27616.1 Artificial Neural Networks......Page 28216.2 Short-Term Load Forecasting......Page 28416.2.1 ANNSTLF Architecture......Page 28516.2.2 Humidity and Wind Speed......Page 28616.2.3 Holidays and Special Days......Page 28716.2.4 Performance......Page 28816.3 Short-Term Price Forecasting......Page 290References......Page 29217.1 Bulk Power System Reliability and Supply Point Reliability......Page 29417.1.1 Bulk Transmission Systems Reliability Is Evaluated Using Deterministic Reliability Criteria......Page 29517.1.2 Supply Point Reliability Is Evaluated Using Either Deterministic or Probabilistic Reliability Criteria......Page 29617.2.1 Reliability Measures—Reliability Indices......Page 29717.2.3 Cost of Interruptions to Consumers......Page 29817.2.4 Outage Models......Page 29917.3.1 Contingency Enumeration Approach......Page 30117.4.1 Calculation of the Reliability of Electric Power Supply to a Major Industrial Complex......Page 30317.4.2 Local Area Reliability......Page 305References......Page 30718 Power System Planning......Page 31018.2 Arenas......Page 31118.3.1 Options......Page 31218.3.2 Uncertainties......Page 31318.3.3 Attributes......Page 31618.4.1 Setting Standards or Criteria......Page 31718.4.5 Least-Cost Planning......Page 31818.4.6 Making Choices......Page 319References......Page 32119 Power System Reliability......Page 32219.1 NERC Regions......Page 32319.2 System Adequacy Assessment......Page 32419.3 System Security Assessment......Page 32519.4 Probabilistic Security Assessment......Page 32619.5 Distribution System Reliability......Page 32719.7 Distribution Reliability Indices......Page 32819.8 Storms and Major Events......Page 33019.9 Component Reliability Data......Page 33119.10 Utility Reliability Problems......Page 33219.11 Reliability Economics......Page 335References......Page 33620.1 Uncertainty in Power System Engineering......Page 33820.2 Deterministic Power Flow Studies......Page 33920.3 Monte Carlo Power Flow Studies......Page 34020.4 Analytical Probabilistic Power Flow Studies......Page 34120.5 Applications for Available Transfer Capability (ATC)......Page 34320.7 An Example of Expected Financial Income from Transmission Tariffs......Page 34520.8 Wind Energy Resources and Stochastic Power Flow Studies......Page 347References......Page 34821.1 Electricity Pricing Overview......Page 35021.2 Electric Cost-of-Service Study......Page 35221.3 Cost-of-Service Study Framework......Page 35321.4 Minimum Distribution System Analysis......Page 35621.5 Analysis of Load Diversity......Page 35921.6 Analysis of Demand and Energy Losses......Page 36321.6.1 Methodology for Evaluating Losses......Page 36721.8 Cost Curve Development......Page 37021.8.1 Coincidence Factor–Load Factor Relationship......Page 37121.8.2 Allocation of Unit Demand Cost Components......Page 37221.9 Rate Design Methodology......Page 375Reference......Page 37822.1 Introduction......Page 38022.2 Accounting......Page 38122.2.2 Amortization and Depreciation......Page 38422.2.3 Financial Statements......Page 38522.3.1 Time Value of Money......Page 38622.4.1 Diversification......Page 38822.4.2 Portfolio Theory......Page 38922.4.3 Capital Asset Pricing Model......Page 39022.4.4 Financial Options......Page 39122.5 Financial Ratios......Page 39322.6 Asset Management......Page 395IV Power Electronics......Page 39823.1 Thyristor and Triac......Page 40023.2 Gate Turn-Off Thyristor......Page 40223.4 Power Bipolar Junction Transistor......Page 40323.5 Power MOSFET......Page 40423.6 Insulated-Gate Bipolar Transistor......Page 40523.7 Integrated Gate-Commutated Thyristor......Page 407Further Information......Page 40824.1 Introduction......Page 41024.2.1 Mechanics of Diode Conduction [4]......Page 41124.2.2 Single-Phase Half-Wave Rectifier Circuits......Page 41224.2.3 Full-Wave Rectifiers [1]......Page 41424.2.4 Three-Phase Rectifiers (Half Wave and Full Wave)......Page 41824.2.5 Average Output Voltage......Page 42024.2.7 Why VFDs Generate Harmonics?......Page 42124.2.8 Harmonic Limit Calculations Based on IEEE 519-1992 [5]......Page 42224.2.10.1 Three-Phase Line Reactors......Page 42424.2.10.2 DC Link Choke......Page 42624.2.10.3 AC Reactor versus DC Link Choke......Page 42724.2.11.2 Shunt Passive Filter......Page 42924.2.11.3 Low-Pass Broadband Filter......Page 43024.2.12.1 12-Pulse Techniques......Page 43124.2.12.2 18-Pulse Techniques......Page 43724.2.12.3 Summary of Drawbacks with Autotransformers......Page 44024.2.12.5 Active Harmonic Compensation......Page 44124.3 Controlled Rectifiers......Page 44224.3.1 Gate Circuit Requirements......Page 44524.3.2 Single-Phase H-Bridge Rectifier Circuits with Thyristors [1–3]......Page 44624.3.3 Three-Phase Controlled AC to DC Rectifier Systems......Page 44924.3.5 Use of Thyristors for Soft Charging DC Bus of Voltage Source Inverters......Page 45124.3.6 HVDC Transmission Systems......Page 45424.3.7 Power System Interaction with Three-Phase Thyristor AC to DC Rectifier Systems......Page 455References......Page 45725.1 Introduction and Overview......Page 45825.3 Single-Phase Inverters......Page 45925.4 Three-Phase Inverters......Page 46025.5 Multilevel Inverters......Page 46425.6 Line-Commutated Inverters......Page 468References......Page 46926 Active Filters for Power Conditioning......Page 47226.1 Harmonic-Producing Loads......Page 47326.2 Theoretical Approach to Active Filters for Power Conditioning......Page 47426.3.1 System Configuration......Page 48126.3.2 Operating Principle......Page 48226.3.3 Control Circuit......Page 48326.3.4 Experimental Results......Page 48426.4.1 Present Status and Future Trends......Page 48626.4.3 The 48-MVA Shunt Active Filter for Compensation of Voltage Impact Drop, Variation, and Imbalance......Page 487References......Page 48927.1 Introduction......Page 49227.2.1 Shunt Compensation Principles......Page 49427.2.2 Traditional Shunt Compensation......Page 49527.2.3 Self-Commutated Shunt Compensators......Page 50127.2.3.1 Principles of Operation......Page 50227.2.3.2 Multilevel Converters......Page 50327.2.3.3 STATCOM Design Principles......Page 50627.2.4 Comparison between Thyristorized and Self-Commutated Compensators......Page 50927.2.5 Superconducting Magnetic Energy Storage......Page 51127.3.1 Series Compensation Principles......Page 51227.3.2 Static Synchronous Series Compensator......Page 51327.3.2.1 Compensation Strategies......Page 51427.3.2.2 Reference Signal Generation......Page 51727.3.2.3 Power Circuit Design......Page 52027.4.1 Unified Power Flow Controller......Page 52327.4.2 Interline Power Flow Controller......Page 52427.4.3 Unified Power Quality Conditioner......Page 52527.5 Facts Controller’s Applications......Page 52627.5.2 Channel Tunnel Rail Link [31]......Page 52727.5.3 STATCOM “Voltage Controller” ±100 MVAr STATCOM at Sullivan Substation (TVA) in Northeastern Tennessee [25]......Page 52827.5.4 Unified Power Flow Controller “All Transmission Parameters Controller”: ±160 MVA Shunt and ±160 MVA Series at Inez Substation (AEP), Northeastern Virginia, USA [25]......Page 52927.5.5 Convertible Static Compensator in the New York 345 kV Transmission System [35]......Page 53027.6 Conclusions......Page 531References......Page 53228.2.1 Basic Concepts of Wind Power Systems......Page 53428.2.2.1 Power Electronics for Wind Turbine Type 1......Page 53528.2.2.2 Power Electronics for Wind Turbine Type 2......Page 53628.2.2.3 Power Electronics for Wind Turbine Type 3......Page 53828.2.3 Control of Power Electronic Converters for Variable-Speed Wind Turbine Generators......Page 54228.2.3.1 Control of Power Electronic Converters for Wind Turbine Type 3......Page 54428.2.3.2 Control of Power Electronic Converters for Wind Turbine Type 4......Page 54828.2.4 Power Electronics for Wind Power Plants......Page 54928.3.1 Basic Concepts of PV Power Systems......Page 55128.3.2 Topologies of PV Power Systems......Page 55228.3.3 Topologies and Control for PV Inverters......Page 554References......Page 555*

The Electric Power Engineering Handbook

POWER SYSTEMS THIRD EDITION

The Electric Power Engineering Handbook Third Edition Edited by

Leonard L. Grigsby

Electric Power Generation, Transmission, and Distribution, Third Edition Edited by Leonard L. Grigsby

Electric Power Transformer Engineering, Third Edition Edited by James H. Harlow

Electric Power Substations Engineering, Third Edition Edited by John D. McDonald

Power Systems, Third Edition Edited by Leonard L. Grigsby

Power System Stability and Control, Third Edition Edited by Leonard L. Grigsby

The Electric Power Engineering Handbook

POWER SYSTEMS THIRD EDITION EDITED BY

LEONARD L. GRIGSBY

Boca Raton London New York

CRC Press is an imprint of the Taylor & Francis Group, an informa business

CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2012 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Version Date: 20111104 International Standard Book Number-13: 978-1-4398-5634-5 (eBook - PDF) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http:// www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com

Contents Preface....................................................................................................................... ix Editor......................................................................................................................... xi Contributors........................................................................................................... xiii

Part I Power System Analysis and Simulation Andrew P. Hanson.................................................................................................................................... I-1

1

The Per-Unit System....................................................................................... 1-1

2

Symmetrical Components for Power System Analysis..................................2-1

3

Power Flow Analysis.......................................................................................3-1

4

Fault Analysis in Power Systems....................................................................4-1

5

Computational Methods for Electric Power Systems.. ...................................5-1

Charles A. Gross

Anthony F. Sleva

Leonard L. Grigsby and Andrew P. Hanson Charles A. Gross

Mariesa L. Crow

Part II Power System Transients Pritindra Chowdhuri.............................................................................................................................. II-1

6

Characteristics of Lightning Strokes..............................................................6-1

7

Overvoltages Caused by Direct Lightning Strokes........................................ 7-1

8

Overvoltages Caused by Indirect Lightning Strokes.....................................8-1

Francisco De la Rosa

Pritindra Chowdhuri Pritindra Chowdhuri

v

vi

Contents

9

Switching Surges.............................................................................................9-1

10

Very Fast Transients.. .................................................................................... 10-1

11

Transmission System Transients: Grounding.............................................. 11-1

12

Transient Recovery Voltage.......................................................................... 12-1

13

Surge Arresters.............................................................................................. 13-1

14

Insulation Coordination............................................................................... 14-1

Stephen R. Lambert

Juan A. Martinez-Velasco William A. Chisholm

Thomas E. McDermott Thomas E. McDermott Stephen R. Lambert

Part III Power System Planning (Reliability) Gerald B. Sheblé.........................................................................................................................................III-1

15

Planning Environments................................................................................ 15-1

16

Short-Term Load and Price Forecasting with Artificial Neural Networks.. ....................................................................................... 16-1

Gerald B. Sheblé

Alireza Khotanzad

17

Transmission Plan Evaluation: Assessment of System Reliability.. ............. 17-1

18

Power System Planning................................................................................. 18-1

19

Power System Reliability.. ............................................................................. 19-1

20

Probabilistic Methods for Planning and Operational Analysis..................20-1

21

Engineering Principles of Electricity Pricing.............................................. 21-1

22

Business Essentials........................................................................................22-1

N. Dag Reppen and James W. Feltes Hyde M. Merrill

Richard E. Brown

Gerald T. Heydt and Peter W. Sauer Lawrence J. Vogt

Richard E. Brown

Contents

vii

Part IV Power Electronics R. Mark Nelms.......................................................................................................................................IV-1

23

Power Semiconductor Devices.. ....................................................................23-1

24

Uncontrolled and Controlled Rectifiers.. .....................................................24-1

25

Inverters.. .......................................................................................................25-1

26

Active Filters for Power Conditioning.. ........................................................26-1

27

FACTS Controllers........................................................................................ 27-1

28

Power Electronics for Renewable Energy.....................................................28-1

Kaushik Rajashekara and Z. John Shen Mahesh M. Swamy

Michael G. Giesselmann Hirofumi Akagi

Luis Morán, Juan Dixon, M. José Espinoza, and José Rodríguez Wei Qiao

Preface The generation, delivery, and utilization of electric power and energy remain one of the most challenging and exciting fields of electrical engineering. The astounding technological developments of our age are highly dependent upon a safe, reliable, and economic supply of electric power. The objective of the Electric Power Engineering Handbook is to provide a contemporary overview of this far-reaching field as well as a useful guide and educational resource for its study. It is intended to define electric power engineering by bringing together the core of knowledge from all of the many topics encompassed by the field. The chapters are written primarily for the electric power engineering professional who seeks factual information, and secondarily for the professional from other engineering disciplines who wants an overview of the entire field or specific information on one aspect of it. The first and second editions of this handbook were well received by readers worldwide. Based upon this reception and the many recent advances in electric power engineering technology and applications, it was decided that the time was right to produce a third edition. Because of the efforts of many individuals, the result is a major revision. There are completely new chapters covering such topics as FACTS, smart grid, energy harvesting, distribution system protection, electricity pricing, linear machines. In addition, the majority of the existing chapters have been revised and updated. Many of these are major revisions. The handbook consists of a set of five books. Each is organized into topical parts and chapters in an attempt to provide comprehensive coverage of the generation, transformation, transmission, distribution, and utilization of electric power and energy as well as the modeling, analysis, planning, design, monitoring, and control of electric power systems. The individual chapters are different from most technical publications. They are not journal-type articles nor are they textbooks in nature. They are intended to be tutorials or overviews providing ready access to needed information while at the same time providing sufficient references for more in-depth coverage of the topic. This book is devoted to the subjects of power system protection, power system dynamics and stability, and power system operation and control. If your particular topic of interest is not included in this list, please refer to the list of companion books referred to at the beginning. In reading the individual chapters of this handbook, I have been most favorably impressed by how well the authors have accomplished the goals that were set. Their contributions are, of course, key to the success of the book. I gratefully acknowledge their outstanding efforts. Likewise, the expertise and dedication of the editorial board and section editors have been critical in making this handbook possible. To all of them I express my profound thanks. They are as follows: • • • •

Nonconventional Power Generation Conventional Power Generation Transmission Systems Distribution Systems

Saifur Rahman Rama Ramakumar George G. Karady William H. Kersting ix

x

Preface

• • • • • • • • • • •

Electric Power Utilization Power Quality Transformer Engineering (a complete book) Substations Engineering (a complete book) Power System Analysis and Simulation Power System Transients Power System Planning (Reliability) Power Electronics Power System Protection Power System Dynamics and Stability Power System Operation and Control

Andrew P. Hanson S. Mark Halpin James H. Harlow John D. McDonald Andrew P. Hanson Pritindra Chowdhuri Gerry Sheblé R. Mark Nelms Miroslav M. Begovic* Prabha S. Kundur† Bruce Wollenberg

I wish to say a special thank-you to Nora Konopka, engineering publisher for CRC Press/Taylor & Francis, whose dedication and diligence literally gave this edition life. I also express my gratitude to the other personnel at Taylor & Francis who have been involved in the production of this book, with a special word of thanks to Jessica Vakili. Their patience and perseverance have made this task most pleasant. Finally, I thank my longtime friend and colleague—Mel Olken, editor, the Power and Energy Magazine—for graciously providing the picture for the cover of this book.

* Arun Phadke for the first and second editions. † Richard Farmer for the first and second editions.

Editor Leonard L. (“Leo”) Grigsby received his BS and MS in electrical engineering from Texas Tech University, Lubbock, Texas and his PhD from Oklahoma State University, Stillwater, Oklahoma. He has taught electrical engineering at Texas Tech University, Oklahoma State University, and Virginia Polytechnic Institute and University. He has been at Auburn University since 1984, first as the Georgia power distinguished professor, later as the Alabama power distinguished professor, and currently as professor emeritus of electrical engineering. He also spent nine months during 1990 at the University of Tokyo as the Tokyo Electric Power Company endowed chair of electrical engineering. His teaching interests are in network analysis, control systems, and power engineering. During his teaching career, Professor Grigsby received 13 awards for teaching excellence. These include his selection for the university-wide William E. Wine Award for Teaching Excellence at Virginia Polytechnic Institute and University in 1980, the ASEE AT&T Award for Teaching Excellence in 1986, the 1988 Edison Electric Institute Power Engineering Educator Award, the 1990–1991 Distinguished Graduate Lectureship at Auburn University, the 1995 IEEE Region 3 Joseph M. Beidenbach Outstanding Engineering Educator Award, the 1996 Birdsong Superior Teaching Award at Auburn University, and the IEEE Power Engineering Society Outstanding Power Engineering Educator Award in 2003. Professor Grigsby is a fellow of the Institute of Electrical and Electronics Engineers (IEEE). During 1998–1999, he was a member of the board of directors of IEEE as the director of Division VII for power and energy. He has served the institute in 30 different offices at the chapter, section, regional, and international levels. For this service, he has received seven distinguished service awards, such as the IEEE Centennial Medal in 1984, the Power Engineering Society Meritorious Service Award in 1994, and the IEEE Millennium Medal in 2000. During his academic career, Professor Grigsby has conducted research in a variety of projects related to the application of network and control theory to modeling, simulation, optimization, and control of electric power systems. He has been the major advisor for 35 MS and 21 PhD graduates. With his students and colleagues, he has published over 120 technical papers and a textbook on introductory network theory. He is currently the series editor for the Electrical Engineering Handbook Series published by CRC Press. In 1993, he was inducted into the Electrical Engineering Academy at Texas Tech University for distinguished contributions to electrical engineering.

xi

Contributors Hirofumi Akagi Department of Electrical and Electronic Engineering Tokyo Institute of Technology Tokyo, Japan Richard E. Brown Quanta Technology Cary, North Carolina William A. Chisholm Kinectrics/Université du Quebec à Chicoutimi Toronto, Ontario, Canada Pritindra Chowdhuri Tennessee Technological University Cookeville, Tennessee Mariesa L. Crow Department of Electrical and Computer Engineering Missouri University of Science and Technology Rolla, Missouri Francisco De la Rosa Electric Power Systems Fremont, California Juan Dixon Department of Electrical Engineering Pontificia Universidad Católica de Chile Santiago, Chile M. José Espinoza Department of Electrical Engineering Universidad de Concepción Concepción, Chile

James W. Feltes Siemens Power Technologies International Schenectady, New York Michael G. Giesselmann Department of Electrical and Computer Engineering Center for Pulsed Power & Power Electronics Texas Tech University Lubbock, Texas Leonard L. Grigsby Department of Electrical and Computer Engineering Auburn University Auburn, Alabama Charles A. Gross Department of Electrical and Computer Engineering Auburn University Auburn, Alabama Andrew P. Hanson The Structure Group Raleigh, North Carolina Gerald T. Heydt School of Electrical, Computer, and Energy Engineering Arizona State University Tempe, Arizona Alireza Khotanzad Department of Electrical Engineering Southern Methodist University Dallas, Texas xiii

xiv

Contributors

Stephen R. Lambert Shawnee Power Consulting, LLC Williamsberg, Virginia

N. Dag Reppen Niskayuna Power Consultants, LLC Niskayuna, New York

Juan A. Martinez-Velasco Department of Electrical Engineering Universitat Politecnica de Catalunya Barcelona, Spain

José Rodríguez Department of Electronics Engineering Universidad Téchnica Federico Santa María Valparaiso, Chile

Thomas E. McDermott Meltran, Inc. Pittsburgh, Pennsylvania

Peter W. Sauer Department of Electrical and Computer Engineering University of Illinois at Urbana-Champaign Champaign, Illinois

Hyde M. Merrill Merrill Energy, LLC Salt Lake City, Utah Luis Morán Department of Electrical Engineering Universidad de Concepción Concepción, Chile R. Mark Nelms Department of Electrical and Computer Engineering Auburn University Auburn, Alabama Wei Qiao Department of Electrical Engineering University of Nebraska-Lincoln Lincoln, Nebraska Kaushik Rajashekara Rolls-Royce Corporation Indianapolis, Indiana

Gerald B. Sheblé Quanta Technology, LLC Raleigh, North Carolina Z. John Shen Department of Electrical Engineering and Computer Science University of Central Florida Orlando, Florida Anthony F. Sleva Altran Solutions Cranbury, New Jersey Mahesh M. Swamy Yaskawa America Incorporated Waukegan, Illinois Lawrence J. Vogt Mississippi Power Company Gulfport, Mississippi

Power System Analysis and Simulation

I

Andrew P. Hanson

1 The Per-Unit System Charles A. Gross...................................................................................1-1

2 Symmetrical Components for Power System Analysis Anthony F. Sleva.......................2-1

3 Power Flow Analysis Leonard L. Grigsby and Andrew P. Hanson.....................................3-1

4 Fault Analysis in Power Systems Charles A. Gross..............................................................4-1

5 Computational Methods for Electric Power Systems Mariesa L. Crow..........................5-1

Impact on Transformers • Per-Unit Scaling Extended to Three-Phase Systems • Per-Unit Scaling Extended to a General Three-Phase System

Introduction • Discussion of Per Unit Quantities • Fundamental Principles • “a” Operator • Phase and Sequence Relationships • Positive Sequence Networks • Sample Three-Phase Fault Calculation: Fault at Bus T2H • Sample Three-Phase Fault Calculation: Fault at Bus T2L • Negative Sequence Networks • Sample Phase to Phase Fault Calculation: Fault at Bus T2L • Zero Sequence Networks • Sample Phase to Ground Fault Calculation: Fault at Bus T2L Introduction • Power Flow Problem • Formulation of Bus Admittance Matrix • Formulation of Power Flow Equations • P–V Buses • Bus Classifications • Generalized Power Flow Development • Solution Methods • Component Power Flows • References • Further Information

Simplifications in the System Model • The Four Basic Fault Types • An Example Fault Study • Further Considerations • Summary • Defining Terms • References • Further Information Power Flow • Optimal Power Flow • State Estimation

I-1

I-2

Power System Analysis and Simulation

Andrew Hanson is a senior manager with The Structure Group, Raleigh, North Carolina. He has 20 years of experience in power system engineering, operations, and consulting, having worked at Tampa Electric, Siemens, and ABB and having led the development of an office for a small privately held engineering firm. His expertise is focused on power delivery system operations and planning, having led the development of processes, forecasts, plans, and distribution automation implementations for a number of utilities. Dr. Hanson has a BEE from The Georgia Institute of Technology and an MEE and a PhD in electrical engineering from Auburn University.

1 The Per-Unit System Charles A. Gross

1.1 1.2 1.3

Auburn University

Impact on Transformers................................................................... 1-4 Per-Unit Scaling Extended to Three-Phase Systems.................... 1-8 Per-Unit Scaling Extended to a General Three- Phase System.....................................................................1-12

In many engineering situations, it is useful to scale or normalize quantities. This is commonly done in power system analysis, and the standard method used is referred to as the per-unit system. Historically, this was done to simplify numerical calculations that were made by hand. Although this advantage has been eliminated by using the computer, other advantages remain: • Device parameters tend to fall into a relatively narrow range, making erroneous values conspicuous. • The method is defined in order to eliminate ideal transformers as circuit components. • The voltage throughout the power system is normally close to unity. Some disadvantages are that component equivalent circuits are somewhat more abstract. Sometimes phase shifts that are clearly present in the unscaled circuit are eliminated in the per-unit circuit. It is necessary for power system engineers to become familiar with the system because of its wide industrial acceptance and use and also to take advantage of its analytical simplifications. This discussion is limited to traditional AC analysis, with voltages and currents represented as complex phasor values. Per-unit is sometimes extended to transient analysis and may include quantities other than voltage, power, current, and impedance. The basic per-unit scaling equation is Per-unit value =

actual value . base value

(1.1)

The base value always has the same units as the actual value, forcing the per-unit value to be dimensionless. Also, the base value is always a real number, whereas the actual value may be complex. Representing a complex value in polar form, the angle of the per-unit value is the same as that of the actual value. Consider complex power

S = VI *

(1.2)

or

S∠θ = V∠αI∠−β

where V = phasor voltage, in volts I = phasor current, in amperes 1-1

1-2

Power Systems

Suppose we arbitrarily pick a value Sbase, a real number with the units of volt-amperes. Dividing through by Sbase, S∠θ V∠αI∠−β = . S base S base

We further define

Vbase Ibase = S base .

(1.3)

Either V base or Ibase may be selected arbitrarily, but not both. Substituting Equation 1.3 into Equation 1.2, we obtain S∠θ V∠α(I∠−β) = S base Vbase Ibase V∠α I∠−β S pu ∠θ = Vbase Ibase

(1.4)

S pu = Vpu ∠α(Ipu ∠−β) S pu = Vpu Ipu*

The subscript pu indicates per-unit values. Note that the form of Equation 1.4 is identical to Equation 1.2. This was not inevitable, but resulted from our decision to relate V base I base and Sbase through Equation 1.3. If we select Z base by Z base =

2 Vbase Vbase = . Ibase S base

(1.5)

Convert Ohm’s law: Z=

V I

(1.6)

into per-unit by dividing by Zbase. Z V/ I = Z base Z base Z pu =

Vpu V/ Vbase = . I / Ibase Ipu

Observe that Z pu =

X R + jX R Z = = + j Z base Z base Z base Z base

Z pu = R pu + jX pu

Thus, separate bases for R and X are not necessary:

Z base = R base = X base

(1.7)

1-3

The Per-Unit System

By the same logic, S base = Pbase = Q base

Example 1.1 (a) Solve for Z, I, and S at Port ab in Figure 1.1a. (b) Repeat (a) in per-unit on bases of V base = 100 V and Sbase = 1000 V. Draw the corresponding per-unit circuit.

Solution (a)

Zab = 8 + j12 − j6 = 8 + j6 = 10 ∠36.9° Ω I=

Vab 100∠0° = = 10∠−36.9° A Z ab 10∠36.9°

S = VI* = (100∠0°)(10∠− 36.9°) * = 1000∠36.9° = 800 + j600 VA P = 800 W Q = 600 var

(b) On bases V base and Sbase = 1000 VA: Z base =

2 2 Vbase (100) = = 10 Ω S base 1000

Ibase =

S base 1000 = = 10 A Vbase 100

Vpu =

100∠0° = 1∠0° pu 100

Z pu =

8 + j12 − j6 = 0.8 + j0.6 pu 10

= 1.0∠36.9° pu Ipu =

Vpu 1∠0° = = 1∠−36.9° pu Z pu 1∠36.9°

S pu = Vpu Ipu * = (1∠0°)(1∠−36.9°)* = 1∠36.9° pu = 0.8 + j0.6 pu

I 100 0° V

(a)

a

8Ω

j12 Ω

+

+

–

1 0° –j6 Ω pu –

b

0.8 pu

I

(b)

j1.2 pu

–j0.6 pu

b

FIGURE 1.1 (a) Circuit with elements in SI units. (b) Circuit with elements in per-unit.

1-4

Power Systems Converting results in (b) to SI units: I = (Ipu )Ibase = (1∠−36.9°)(10) = 10∠−36.9° A Z = (Z pu )Z base = (0.8 + j0.6)(10) = 8 + j6 Ω S = (S pu )S base = (0.8 + j0.6)(1000) = 800 + j600 W, var

The results of (a) and (b) are identical.

For power system applications, base values for Sbase and V base are arbitrarily selected. Actually, in practice, values are selected that force results into certain ranges. Thus, for V base, a value is chosen such that the normal system operating voltage is close to unity. Popular power bases used are 1, 10, 100, and 1000 MVA, depending on system size.

1.1 Impact on Transformers To understand the impact of pu scaling on transformer, consider the three-winding ideal device (see Figure 1.2). For sinusoidal steady-state performance:

V1 =

N1 V2 N2

(1.8a)

V2 =

N2 V3 N3

(1.8b)

V3 =

N3 V1 N1

(1.8c)

and N1I1 + N2I2 + N3I3 = 0

(1.9)

+

I1

I2

V2

+

– V1

I3 +

–

V3 – N 1 : N2 : N 3 Ideal

FIGURE 1.2 The three-winding ideal transformer.

1-5

The Per-Unit System

Consider the total input complex power S. S = V1I1* + V2I2* + V3I3* = V1I1* + =

N2 N V1I2* + 3 V1I3* N1 N1

V1 [N1I1 + N2I2 + N3I3 ]* N1

=0

(1.10)

The interpretation to be made here is that the ideal transformer can neither absorb real nor reactive power. An example should clarify these properties. Arbitrarily select two base values V1base and S1base. Require base values for windings 2 and 3 to be

V2 base =

N2 V1base N1

(1.11a)

V3base =

N3 V1base N1

(1.11b)

and

S1base = S2 base = S3base = S base

(1.12)

By definition,

I1base =

S base V1base

(1.13a)

I2base =

S base V2base

(1.13b)

I3base =

S base V3base

(1.13c)

I2 base =

N1 I1base N2

(1.14a)

I3base =

N1 I1base N3

(1.14b)

It follows that

Recall that a per-unit value is the actual value divided by its appropriate base. Therefore:

V1 (N / N ) V = 1 2 2 V1base V1base

(1.15a)

1-6

Power Systems

and

(N1/ N2 )V2 V1 = V1base (N1/ N2 )V2 base

(1.15b)

V1pu = V2 pu

(1.15c)

(N1/ N3 )V3 V1 = V1base (N1/ N3 )V3base

(1.16a)

V1pu = V3pu

(1.16b)

V1pu = V2 pu = V3pu

(1.17)

or indicates per-unit values. Similarly,

or Summarizing: Divide Equation 1.9 by N1

I1 +

N2 N I 2 + 3 I3 = 0 N1 N1

Now divide through by I1base (N / N )I (N / N )I I1 + 2 1 2 + 3 1 3 =0 I1base I1base I1base I1

I1base

+

(N3 / N1 )I3 (N2 / N1 )II2 + =0 (N2 / N1 )I2 base (N3 / N1 )I3base

Simplifying to I1pu + I2 pu + I3pu = 0

(1.18)

Equations 1.17 and 1.18 suggest the basic scaled equivalent circuit, shown in Figure 1.3. It is cumbersome to carry the pu in the subscript past this point: no confusion should result, since all quantities will show units, including pu.

I1

+ V2

1

+ V1

I2

2

–

FIGURE 1.3 Single-phase ideal transformer.

3

– + –

V3

I3

1-7

The Per-Unit System 2 I1 + 1 0°

+

1

V2

+ V1

–

I2 = 3.5 –30°

–

3

I3 = 0.5 +30°

+ V3 – –

FIGURE 1.4 Per-unit circuit.

Example 1.2 The three-winding single-phase transformer of Figure 1.1 is rated at 13.8 kV/138 kV/4.157 kV and 50 MVA/40 MVA/10 MVA. Terminations are as follows: 13.8 kV winding: 13.8 kV Source 138 kV winding: 35 MVA load, pf = 0.866 lagging 4.157 kV winding: 5 MVA load, pf = 0.866 leading Using Sbase = 10 MVA, and voltage ratings as bases,

(a) Draw the pu equivalent circuit. (b) Solve for the primary current, power, and power, and power factor. Solution

(a) See Figure 1.4. (b) S2 =

35 = 3.5 pu S2 = 3.5∠+ 30° pu 10

S3 =

5 = 0.5 pu S3 = 0.5∠− 30° pu 10

V1 =

13.8 = 1.0 pu V1 = V2 = V3 = 1.0∠0° pu 1 3. 8

S I2 = 2 * = 3.5∠− 30° pu V2

S I3 = 3 * = 0.5∠+ 30° pu V3 All values in Per-Unit Equivalent Circuit: I1 = I2 + I3 = 3.5∠−30° + 0.5∠+30° = 3.464 − j1.5 = 3.775∠−23.4° pu S1 = V1I1 * = 3.775∠+23.4° pu S1 = 3.775(10) = 37.75 MVA; pf = 0.9177 lagging

10 = 2736 A I1 = 3.775 0.0138

1-8

Power Systems

1.2 Per-Unit Scaling Extended to Three-Phase Systems The extension to three-phase systems has been complicated to some extent by the use of traditional terminology and jargon, and a desire to normalize phase-to-phase and phase-to-neutral voltage simultaneously. The problem with this practice is that it renders Kirchhoff’s voltage and current laws invalid in some circuits. Consider the general three-phase situation in Figure 1.5, with all quantities in SI units. Define the complex operator: a = 1∠120°

The system is said to be balanced, with sequence abc, if Vbn = a 2 Van Vcn = aVan

and I b = a 2 Ia Ic = aIa −In = Ia + Ib + Ic = 0

Likewise:

Vab = Van − Vbn Vbc = Vbn − Vcn = a2 Vab Vca = Vcn − Van = aVab

If the load consists of wye-connected impedance: Zy =

Van Vbn Vcn = = Ia Ib Ic

The equivalent delta element is Z ∆ = 3Z Y

a General source

b

+

Ib Van

c n

FIGURE 1.5 General three-phase system.

Ia

–

+ –

Vbn

Ic Vcn In

General load

1-9

The Per-Unit System

To convert to per-unit, define the following bases: S3ϕbase = The three-phase apparent base at a specific location in a three-phase system, in VA. VLbase = The line (phase-to-phase) rms voltage base at a specific location in a three-phase system, in V. From the above, define: S base =

Vbase =

S3φbase 3 VLbase 3

(1.19)

(1.20)

It follows that

Ibase =

S base Vbase

(1.21)

Z base =

Vbase Ibase

(1.22)

An example will be useful. Example 1.3 Consider a balanced three-phase 60 MVA 0.8 pf lagging load, sequence abc operating from a 13.8 kV (line voltage) bus. On bases of S3ϕbase = 100 MVA and VLbase = 13.8 kV:

(a) Determine all bases. (b) Determine all voltages, currents, and impedances in SI units and per-unit. Solution

(a)

S base =

Vbase =

VLbase 13.8 = = 7.967 kV 3 3

Ibase =

S base = 4.184 kA Vbase

Z base =

S3φbase 100 = = 33.33 MVA 3 3

Vbase = 1.904 Ω Ibase

(b)

Van = 7.967 ∠0° kV (1.000∠0° pu) Vbn = 7.967 ∠−120° kV (1.000∠−120° pu) Vcn = 7.967 ∠+120° kV (1.000∠+120° pu)

1-10

Power Systems

Sa = S b = S c =

S3φ 60 = = 20 MVA(0.60 pu) 3 3

Sa = S b = Sc = 16 + j12MVA(0.48 + j0.36 pu)

S Ia = a = 2.510∠− 36.9° kA(0.6000∠− 36.9° pu) Van

Ib = 2.510∠−156.9° kA(0.6000∠−156.9° pu)

Ic = 2.510∠83.1° kA(0.6000∠83.1° pu)

ZY =

Van = 3.174 ∠+ 36.9° = 2.539 + j1.904 Ω(1.33 + j1.000 pu) Ia Z ∆ = 3ZY = 7.618 + j5.713 Ω(4 + j3 pu)

Vab = Van − Vbn = 13.8∠30° kV(1.732∠30° pu)

Vbc = 13.8∠− 90° kV(1.732∠90° pu)

Vca = 13.8∠150° kV(1.732∠150° pu)

Converting voltages and currents to symmetrical components:

1 V0 1 V1 = 1 3 V2 1

1 a a2

1 Van 0 kV a 2 Vbn = 7.967 ∠0° kV a Vcn 0 kV

(0 pu) (1∠0° pu) (0 pu)

I0 = 0 kA(0 pu) I1 = 2.510∠− 36.9° kA(0.6∠− 36.9° pu) I2 = 0 kA(0 pu)

Inclusion of transformers demonstrates the advantages of per-unit scaling.

Example 1.4 A 3ϕ 240 kV :15 kV transformer supplies a 13.8 kV 60 MVA pf = 0.8 lagging load, and is connected to a 230 kV source on the high-voltage (HV) side, as shown in Figure 1.6.

(a) Determine all base values on both sides for S3ϕbase = 100 MVA. At the low-voltage (LV) bus, VLbase = 13.8 kV. (b) Draw the positive sequence circuit in per-unit, modeling the transformer as ideal. (c) Determine all currents and voltages in SI and per-unit.

1-11

The Per-Unit System Load

HV bus

LV bus

FIGURE 1.6 A three-phase transformer situation.

Solution

(a) Base values on the LV side are the same as in Example 1.3. The turns ratio may be derived from the voltage ratings ratios: N1 240/ 3 = = 16 N2 15/ 3 ∴(Vbase )HV side = (Ibase )HV side =

S base 33.33 = = 261.5 A (Vbase )HV side 0.1275

Results are presented in the following chart.

Bus LV HV

N1 (Vbase )LV side = 16.00(7.967) = 127.5 kV N2

S3ϕbase

VL base

Sbase

Ibase

Vbase

Zbase

MVA 100 100

kV 13.8 220.8

MVA 33.33 33.33

kA 4.184 0.2615

kV 7.967 127.5

Ohm 1.904 487.5

(b)

VLV =

7.967 ∠0° = 1∠0° pu 7.967

S1φ =

60 = 20 MVA 3

S1φ =

20 = 0.6 pu 33.33

The positive sequence circuit is shown as Figure 1.7. I = 0.60 –36.9° 1 0°

+ – Transformer

FIGURE 1.7 Positive sequence circuit.

1-12

Power Systems

(c) All values determined in pu are valid on both sides of the transformer! To determine SI values on the HV side, use HV bases. For example: Van = (1∠0°)127.5 = 127.5∠0° kV Vab = (1.732∠30°)(127.5) = 220.8∠30° kV

Ia = (0.6∠− 36.9°)(261.5) = 156.9∠− 36.9° A

Example 1.5 Repeat the previous example using a 3ϕ 240 kV:15 kV Δ Solution All results are the same as before. The reasoning is as follows. The voltage ratings are interpreted as line (phase-to-phase) values independent of connection (wye or delta). Therefore the turns ratio remains: N1 240/ 3 = = 16 N2 15/ 3

As before:

(Van )LV side = 7.967 kV (Van )HV side = 127.5 kV

However, Van is no longer in phase on both sides. This is a consequence of the transformer model, and not due to the scaling procedure. Whether this is important depends on the details of the analysis.

1.3 Per-Unit Scaling Extended to a General Three-Phase System The ideas presented are extended to a three-phase system using the following procedure.

1. Select a three-phase apparent power base (S3ph base), which is typically 1, 10, 100, or 1000 MVA. This base is valid at every bus in the system. 2. Select a line voltage base (VL base), user defined, but usually the nominal rms line-to-line voltage at a user-defined bus (call this the “reference bus”). 3. Compute S base =

(S3ph base ) 3

(Valid at every bus)

(1.23)

4. At the reference bus: Vbase =

Ibase =

VL base 3 S base Vbase

(1.24)

(1.25)

1-13

The Per-Unit System 2

Z base =

Vbase Vbase = I base S base

(1.26)

5. To determine the bases at the remaining busses in the system, start at the reference bus, which we will call the “from” bus, and execute the following procedure: Trace a path to the next nearest bus, called the “to” bus. You reach the “to” bus by either passing over (a) a line or (b) a transformer. a. The “line” case: VL base is the same at the “to” bus as it was at the “from” bus. Use Equations 1.2 through 1.4 to compute the “to” bus bases. b. The “transformer” case: Apply VL base at the “from” bus, and treat the transformer as ideal. Calculate the line voltage that appears at the “to” bus. This is now the new VL base at the “to” bus. Use Equations 1.2 through 1.4 to compute the “to” bus bases. Rename the bus at which you are located, the “from” bus. Repeat the above procedure until you have processed every bus in the system. 6. We now have a set of bases for every bus in the system, which are to be used for every element terminated at that corresponding bus. Values are scaled according to per - unit value = actual value/base value where actual value = the actual complex value of S, V, Z, or I, in SI units (VA, V, Ω, A); base value = the (user-defined) base value (real) of S, V, Z, or I, in SI units (VA, V, Ω, A); per-unit value = the per-unit complex value of S, V, Z, or I, in per-unit (dimensionless).

Finally, the reader is advised that there are many scaling systems used in engineering analysis, and, in fact, several variations of per-unit scaling have been used in electric power engineering applications. There is no standard system to which everyone conforms in every detail. The key to successfully using any scaling procedure is to understand how all base values are selected at every location within the power system. If one receives data in per-unit, one must be in a position to convert all quantities to SI units. If this cannot be done, the analyst must return to the data source for clarification on what base values were used.

2 Symmetrical Components for Power System Analysis

Anthony F. Sleva Altran Solutions

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12

Introduction....................................................................................... 2-1 Discussion of Per Unit Quantities................................................... 2-3 Fundamental Principles....................................................................2-6 “a” Operator........................................................................................2-6 Phase and Sequence Relationships.................................................. 2-7 Positive Sequence Networks........................................................... 2-10 Sample Three-Phase Fault Calculation: Fault at Bus T2H............2-10 Sample Three-Phase Fault Calculation: Fault at Bus T2L.............2-11 Negative Sequence Networks......................................................... 2-11 Sample Phase to Phase Fault Calculation: Fault at Bus T2L.........2-12 Zero Sequence Networks................................................................ 2-13 Sample Phase to Ground Fault Calculation: Fault at Bus T2L..........................................................................2-16

2.1 Introduction Three-phase power systems are difficult to model and analyze using traditional methods, that is, using loop equations (Kirchoff’s laws) and node equations (Norton’s laws) when transformer turns ratios and transformer winding connections are included in the calculations. In this chapter, an alternate approach using “symmetrical components” will be introduced. The concept was first proposed by C.L. Fortescue in 1918 in a classic paper devoted to consideration of the general N-phase case. Per phase analysis can be developed for the simple three-phase system illustrated in Figure 2.1 by solving Equations 2.1a through 2.1d. After line and residual currents are calculated, voltage can be calculated:

VA − IA × (Z A Source + Z A System + Z A Load) − IR × Z R System = 0

(2.1a)

VB − IB × (Z B Source + Z B System + Z B Load) − IR × Z R System = 0

(2.1b)

VC − IC × (Z C Source + Z C System + Z C Load) − IR × Z R System = 0

(2.1c)

I A + I B + IC − I R = 0

(2.1d)

Solving Equations 2.1 is tedious because all values are vectorial, that is, voltages are represented as a voltage magnitude with an associated angle and impedances are represented as complex impedances that include resistance and reactance components. In addition, the residual system impedance (ZR) is the parallel combination of neutral and earth impedance. 2-1

2-2

Power Systems

VA VB VC

ZA Source

ZA System

IA

ZA Load

ZB Source

ZB System

IB

ZB Load

ZC Source

ZC System

IC

ZC Load

ZR System

IR

FIGURE 2.1 General three-phase circuit.

Using symmetrical components, analysis is greatly simplified because calculations are a function of load or fault condition rather than a function of source voltage. Conversion from phase quantities to symmetrical quantities or from symmetrical quantities to phase quantities is easily accomplished using rules established by Dr. Fortescue. The basic premise underlying symmetrical components is that calculations, for any load or fault condition, can be developed using single phase methods. Figures 2.2 through 2.5 illustrate impedance networks used to calculate three-phase faults, phase to phase faults, single phase to ground faults, and double phase to ground faults. RB1 V1

Z1 Source

Z1 System

FB1

I1

FIGURE 2.2 Three-phase fault—symmetrical component equivalent circuit. RB1

Z1 Source

V1

RB2

Z2 Source

Z1 System

I1

FB1

I2 FB2

Z2 System

FIGURE 2.3 Phase to phase fault—symmetrical component equivalent circuit. RB1

V1

Z1 Source

RB1 Z2 Source

RB0 Z0 Source

I1

FB1

Z2 System I2

FB2

Z1 System

Z0 System I0

FB0

FIGURE 2.4 Single phase to ground fault—symmetrical component equivalent circuit.

2-3

Symmetrical Components for Power System Analysis

V1

Z1 Source

Z2 Source

Z0 Source

Z1 System

I1

Z2 System

I2

Z0 System

I0

FIGURE 2.5 Double phase to ground fault—symmetrical component equivalent circuit.

Generally, two simplifying assumptions are made, namely, source voltages and system impedances are balanced. Although symmetrical components can be used to develop calculations for unbalanced source voltages and/or unbalanced system impedances, such calculations are beyond the scope of this book.

2.2 Discussion of Per Unit Quantities Calculations are developed using per unit quantities for voltage, current, and impedance as per unit quantities eliminate the need to use transformer turns ratios and conversions throughout calculations. The use of per unit quantities, however, introduces the need to convert all impedances to a common MVA base before developing calculations and the need to convert per unit quantities to volts and amps after calculations have been developed. Common practice is to develop calculations using nominal system voltages as base voltages. Calculations with transformers operating with off-nominal tap positions are only developed when special considerations arise. Transformer impedances, which are listed in percent on transformer nameplates, are determined by test during factory acceptance tests. Per unit values are equal to percent values divided by 100. Transformer winding connections need to be included as essential system data. Generator impedances, which are listed on data sheets in per unit quantities, are a function of fault duration. For short circuit analysis, X d″ (generator subtransient reactance) is used. X d′ (generator transient reactance a few cycles after a three-phase fault occurs on the terminals of a generator) and Xd (generator synchronous reactance when a three-phase fault persists on the terminals of a generator) are used when developing specific calculations to analyze generator parameters, generating station circuit breaker interrupting capability, stuck circuit breaker conditions at generating stations, etc. Line (feeder or circuit) impedances are a function of conductor length, size, type, spacing, phase relationship, construction, bundling, etc., are application specific, and need to be calculated for each specific line. (In many applications, standardized line impedances are available and the calculation becomes a function of line length only.) Motor impedance is a function of the condition being analyzed—running, starting, or backfeeding a fault. The duration of motor backfeed is a function of inertia of the driven load. When specific motor data is not available, a common practice is to assume running power factor is 85%, starting power factor is 30%, and motor impedance during backfeed conditions is 0.10 per unit. When developing calculations using per unit quantities, any MVA base can be used. Common practice is to use 1 MVA as the base for low-voltage systems, 10 MVA for medium voltage systems, and 100 MVA for high-voltage systems. After base MVA and base voltage are selected, base amps and

2-4

Power Systems

base ohms are calculated using the Equations 2.2a and 2.2b for three-phase systems and Equations 2.3a and 2.3b for single phase systems: Base amps =

base MVA(3ϕ) × 1000 3 × base kV(ϕ − ϕ)

Base amps =

(2.2b)

base MVA(1ϕ) × 1000 (single phase) base kV(ϕ − n)

(2.3a)

Base ohms =

(2.2a)

(base kV(ϕ − ϕ))2 (three phase) MVA(3ϕ)

Base ohms =

(three phase)

(base kV(ϕ − n))2 MVA(1ϕ)

(single phase)

(2.3b)

Calculated values for base amps and base ohms, when base MVA and base voltage are selected, are shown in Table 2.1. Per unit impedances for components are converted from nameplate values to base values using Equations 2.4a and 2.4b: Impedance at base MVA = Impedance at given MVA × Impedance at base kV = Impedance at given kV ×

base MVA given MVA

given kV 2 base kV 2

Table 2.2 Lists per unit impedances that are converted from given MVA to base MVA. Table 2.3 Lists per unit impedance that are converted from given kV to base kV. TABLE 2.1 Example Base Values Base MVA (MVA)

Base Voltage

Base Amps (A)

Base Ohms (Ω)

1 1 1 1 10 10 10 100 100 100 100 100

120 V, 1φ 277 V, 1φ 208 V, 3φ 480 V, 3φ 12.47 kV, 3φ 13.8 kV, 3φ 34.5 V, 3φ 69 kV, 3φ 138 kV, 3φ 230 kV, 3φ 345 kV, 3φ 500 kV, 3φ

8333 3610 2775 1202 463 418 167 837 418 251 167 115

0.014 0.077 0.043 0.230 15.55 19.0 119 47.6 190 529 1190 2500

(2.4a)

(2.4b)

2-5

Symmetrical Components for Power System Analysis TABLE 2.2 Impedance Conversion to Common MVA Base Given MVA (MVA) 0.25 0.5 1.0 2.0 2.5 5.0 10 20 25 50 100 200

Base MVA (MVA)

Given Impedance (%)

Impedance at Base MVA (%)

1 1 1 1 10 10 10 10 100 100 100 100

1 1 1 1 5 5 5 5 10 10 10 10

4.00 2.00 1.00 0.50 20.00 10.00 5.00 2.50 40.00 20.00 10.00 5.00

TABLE 2.3 Impedance Conversion to Common Voltage Base Given kV (kV) 13.2 13.8 14.1 220 230 245

Base kV (kV)

Given Impedance (%)

Impedance at Base kV (%)

13.8 13.8 13.8 230 230 230

5 5 5 10 10 10

4.57 5.00 5.22 9.15 10.00 11.35

Per unit values can be converted to actual values using the following equations:

Circuit amps = per unit amps × base amps

(2.5a)

Circuit volts = per unit volts × base volts

(2.5b)

Circuit ohms = per unit ohms × base ohms

(2.5c)

Impedance conversions to common voltage and MVA bases are necessary when transformers with different voltage ratings, but similar turns ratios, are connected in parallel. This is illustrated in Table 2.4. TABLE 2.4 Transformer Data Converted to Common Base Transformer 1 Data 25 69 13.2 7% 5.227 190.44 13.33 28.0%

MVA kV kV Impedance Transformer turns ratio Base ohms, 25 MVA, 69 kV Transformer ohms At 100 MVA, 69 kV

Transformer 2 Data 25 66 12.47 7% 5.293 174.24 12.20 25.6%

MVA kV kV Impedance Transformer turns ratio Base ohms, 25 MVA, 66 kV Transformer ohms At 100 MVA, 69 kV

2-6

Power Systems

2.3 Fundamental Principles The fundamental principles of symmetrical components are as follows:

1. Single phase methods can be used to solve problems on three-phase systems. 2. Sequence networks have been developed to calculate sequence voltages and sequence currents. These networks are designated the positive sequence network, the negative sequence network, and the zero sequence network. 3. Each sequence has its own line conductors, transformer equivalent circuits, generator equivalent circuits, etc. 4. Each sequence is independent of the other sequences (for balanced fault conditions). 5. Per unit quantities are used in calculations. 6. Single phase to ground faults are calculated as Aφ to ground faults. 7. Phase to phase faults are calculated as Bφ to Cφ faults. 8. Sequence currents are calculated at a fault bus. Current dividers are used to calculate sequence currents at other points in each sequence network. 9. Sequence voltages are calculated at a fault bus. Voltage dividers are used to calculate sequence voltages at other points in each sequence network. 10. Positive sequence voltage is maximum at a generator and minimum at a fault bus. 11. Negative sequence voltage is maximum at a fault bus and zero at the reference bus. 12. Zero sequence voltage is maximum at a fault bus and zero at the reference bus. 13. Phase voltage can be calculated when sequence voltages are known. 14. Phase current can be calculated when sequence currents are known. 15. Sequence voltages can be calculated if phase voltages are known. 16. Sequence currents can be calculated if phase currents are known.

2.4 “a” Operator Symmetrical components utilize a unit vector, designated “a,” to transform phase quantities to sequence quantities and to transform sequence quantities to phase quantities. The properties of the “a” vector are as follows:

a 0 = 1 @ 0°

(2.6a)

a1 = 1 @ 120° = −0.5 + j0.866

(2.6b)

a 2 = 1 @ 240° = −0.5 − j0.866

(2.6c)

a 3 = 1 @ 360° = a

(2.6d)

a 2 − a = 1 @ 240° − 1 @ 120° = − j1.732 = − j √ 3 a − a 2 = 1 @ 120° − 1 @ 240° = + j1.732 = j √ 3 1 + a + a 2 = 1 + 1 @ 120° + 1 @ 240° = 0

(2.6e) (2.6f) (2.6g)

Symmetrical Components for Power System Analysis

2-7

2.5 Phase and Sequence Relationships Power system voltages and currents are represented as the sum of sequence values:

Ea = Ea1 + Ea2 + Ea0

(2.7a)

E b = E b1 + E b2 + E b0

(2.7b)

Ec = Ec1 + Ec2 + Ec0

(2.7c)

Ia = Ia1 + Ia2 + Ia0

(2.8a)

Ib = Ib1 + Ib2 + Ib0

(2.8b)

Ic = Ic1 + Ic2 + Ic0

(2.8c)

Sequence values are derived from phase voltages and current using the following:

1 Ea1 = (Ea + aE b + a 2Ec ) 3

(2.9a)

1 Ea2 = (Ea + a 2E b + aEc ) 3

(2.9b)

1 Ea 0 = (Ea + E b + Ec ) 3

(2.9c)

1 Ia1 = (Ia + aIb + a 2Ic ) 3

(2.10a)

1 Ia2 = (Ia + a 2Ib + aIc ) 3

(2.10b)

1 Ia 0 = (Ia + Ib + Ic ) 3

(2.10c)

Positive sequence voltages and currents are represented by vectors that have equal amplitudes and are displaced 120° relative to each other as shown in Figure 2.6:

Ea1 = Ea1

(2.11a)

E b1 = a 2Ea1 = Ea1 @ 240°

(2.11b)

Ec1 = aEa1 = Ea1 @ 120°

(2.11c)

2-8

Power Systems Ec1

Ea1

Eb1

FIGURE 2.6 Positive sequence voltage vectors.

Ia1 = Ia1

(2.12a)

Ib1 = a 2Ia1 = Ia1 @ 240°

(2.12b)

Ic1 = aIa1 = Ia1 @ 120°

(2.12c)

Negative sequence voltages and currents are represented by vectors that have equal amplitudes and are displaced 120° relative to each other, as shown in Figure 2.7. The rotation of the negative sequence vectors is opposite the rotation of the positive sequence vectors:

Ea2 = Ea2

(2.13a)

E b2 = aEa2 = Ea2 @ 120°

(2.13b)

Ec2 = a 2Ea2 = Ea2 @ 240°

(2.13c)

Ia2 = Ia2

(2.14a)

Ib2 = aIa2 = Ia2 @ 120°

(2.14b)

Ic2 = a 2Ia2 = Ia2 @ 240°

(2.14c)

Eb2

Ea2

Ec2

FIGURE 2.7 Negative sequence voltage vectors.

2-9

Symmetrical Components for Power System Analysis Ea0 Eb0 Ec0

FIGURE 2.8 Zero sequence voltage vectors.

Zero sequence voltages and currents are represented by vectors that have equal amplitudes that are in phase with each other, as shown in Figure 2.8:

E a0 = E a0

(2.15a)

E b 0 = Ea 0

(2.15b)

E c 0 = Ea 0

(2.15c)

Ia0 = Ia0

(2.16a)

Ib0 = Ia0

(2.16b)

Ic0 = Ia0

(2.16c)

When calculations are developed, sequence quantities, that is, Ia0, Ia1, Ia2 and Ea0, Ea1, Ea2, are calculated. Then, phase quantities are calculated using previous equations. The use of symmetrical components and the determination of the magnitude and angle of positive sequence, negative sequence, and zero sequence vectors for various fault conditions will be demonstrated for the power system shown in Figure 2.9. This power system consists of a generator, two transformers, and one transmission line. For this generator X/R = 40 and for these transformers X/R = 25. When X/R is 10 or less, resistance is generally included in calculations.

Generator 1 50 MVA, 23 kV Xd˝ = 18.5% Xd΄ = 52.3% Xd = 104% X2 = 18.7% X0 = 23.7%

Bus 1

69 kV transmission line Z1L = 1.0 + j9.0 Ω Z0L = 4.1 + j34.7 Ω

Transformer 1 50 MVA, 23–69 kV X = 6.76% 23 kV: ∆ 69 kV: Y-Grd Xd˝ = Generator subtransient reactance Xd΄ = Generator transient reactance Xd = Generator synchronous reactance

FIGURE 2.9 Power system single line diagram.

Transformer 2 10 MVA, 69–13.8 kV X = 7.2% 69 kV: ∆ 13.8 kV: Y-Grd

2-10

Power Systems Bus G1 RB

Z1 Gen 1

Bus T2H

Bus 1

Z1 Trans 1

Z1 Line

Bus T2L

Z1 Trans 2

FIGURE 2.10 Positive sequence network for power system shown in Figure 2.9.

Bus G1 RB1

Bus 1

Z1 Trans 1

Z1 Gen 1 0.009 + j0.370 pu

Bus T2H Z1 Line

Bus T2L

Z1 Trans 2

0.005 + j0.135 pu 0.021 + j0.189 pu 0.029 + j0.720 pu

Base MVA = 100 MVA Base voltages = 23, 69, and 13.8 kV

FIGURE 2.11 Positive sequence network with normalized impedances.

2.6 Positive Sequence Networks Positive sequence networks are in all respects identical to the usual single line diagram for three-phase networks. The resistances and reactances are the values used in traditional equivalent circuits. Although each machine (generator or motor) is considered to be a voltage (power) source, motors are omitted when fault contributions from motors are negligible. Figure 2.10 is the positive sequence network for the system shown in Figure 2.9. The positive sequence reference bus, RB1, the common point to which all power sources are connected, is needed for network analysis. Other buses shown in Figure 2.10 (G1, 1, T2H and T2L) are locations where faults can be analyzed by inserting a jumper from a fault bus (G1, 1, T2H or T2L), to a reference bus, RB1. Positive sequence jumper connections vary with fault type as shown in Figures 2.2 through 2.5. Figure 2.11 is the positive sequence network with normalized impedances. To calculate a three-phase fault, a jumper is placed between RB1 and the fault bus.

2.7 Sample Three-Phase Fault Calculation: Fault at Bus T2H

V1 = 1.00 pu Z1 = 0.036 + j0.694 pu = 0.695 pu @ 87.06° Ia1 = 1.439 pu @ − 87.06° Ia1 = 1204 A @ − 87.06° at 69 kV IA = Ia1 = 1204 A @ − 87.06° IB = a 2Ia1 = 1 @ 240° × 1204 A @ − 87.06° = 1204 A @ 152.94° IC = aIa1 = 1 @ 120° × 1204 A @ − 87.06° = 1204 A @ 32.94°

Symmetrical Components for Power System Analysis

2-11

V1 @ Gen 1 = 1.00 pu − 1.439 pu @ − 87.06° × 0.370 pu @ 90° V1 @ Gen 1 = 0.469 pu @ − 3.34° VA − N @ Gen 1 = 6.23 kV @ − 3.34°

2.8 Sample Three-Phase Fault Calculation: Fault at Bus T2L

V1 = 1.00 pu Z1 = 0.064 + j1.414 pu = 1.416 pu @ 87.39° Ia1 = 0.706 pu @ − 87.39° Ia1 = 2955 A @ − 87.39° at 13.8 kV IA = Ia1 = 2955 A @ − 87. 39° IB = a 2Ia1 = 1 @ 240° × 2955 A @ − 87.39° = 2955 A @ 152.61° IC = aIa1 = 1 @ 120° × 2955 A @ − 87.39° = 2955 A @ 32.61° V1 @ Gen 1 = 1.00 pu − 0.706 pu @ − 87.39° × 0.370 pu @ 90° V1 @ Gen 1 = 0.739 pu @ − 0.92° VA − N @ Gen 1 = 9.81 kV @ − 0.92°

2.9 Negative Sequence Networks Negative sequence networks are similar to positive sequence networks in that the number of branches is the same. But, negative sequence networks do not contain voltage sources because machines (generators or motors) only generate positive sequence voltage. The other difference between negative sequence networks and positive sequence networks is that different machine impedances may be used in each sequence. For most calculations, machine negative sequence impedance is assumed to be equal to positive sequence impedance. Different positive sequence and negative sequence impedances are used when generator response to short circuit conditions or circuit breaker interrupting capabilities are being evaluated at generating stations. Figure 2.12 is the negative sequence network for the system shown in Figure 2.9. The negative sequence reference bus, RB2, the common point to which all power sources are connected, is needed for network analysis. Other buses shown in Figure 2.12 (G1, 1, T2H and T2L) are locations where faults can be analyzed by inserting a jumper from a fault bus to a reference bus. Negative sequence jumper connections vary with fault type, as shown in Figures 2.3 through 2.5.

2-12

Power Systems Bus G1 RB2

Z2 Gen 1

Bus 1

Z2 Trans 1

Bus T2H Z2 Line

Bus T2L

Z2 Trans 2

FIGURE 2.12 Negative sequence network for power system shown in Figure 2.9. Bus G1

RB2

Z2 Gen 1

Bus 1

Z2 Trans 1

0.009 + j0.370 pu

Bus T2H

Z2 Line

Bus T2L

Z2 Trans 2

0.005 + j0.135 pu 0.021 + j0.189 pu 0.029 + j0.720 pu

Base MVA = 100 MVA Base voltages = 23, 69, and 13.8 kV

FIGURE 2.13 Negative sequence network with normalized impedances.

Figure 2.13 is the negative sequence network with normalized impedances. To calculate a phase to phase fault, the positive and negative sequence networks must be connected, as shown in Figure 2.3. This means that one jumper is needed between RB1 and RB2 and a second jumper is needed between the positive sequence fault bus and the corresponding negative sequence fault bus.

2.10 Sample Phase to Phase Fault Calculation: Fault at Bus T2L

V1 = 1.00 pu

Z = Z1 + Z 2 = 2 × (0.064 + j1.414) pu = 2.831 pu @ 87.39°

Ia1 = 0.353 pu @ − 87.39° Ia2 = −Ia1 = 0.353 pu @ 92.61° Ia1 = 1477 A @ − 87.39° at 13.8 kV Ia2 = 1477 A @ 92.61° at 13.8 kV IA = Ia1 + Ia2 = 1477 A @ − 87.39° + 1477 A @ 92.61° at 13.8 kV =0 IB = a 2Ia1 + aIa2 = 1 @ 240° × 1477 A @ − 87. 39° + [email protected] 120° × 1477 A @ 92.61° = 2559 A @ −182.61° IC = aIa1 + a 2Ia2 = [email protected] 120° × 1477 A @ − 87. 39° + [email protected] 240° × 1477 A @ − 92.61° = 2559 A @ − 2.61°

2-13

Symmetrical Components for Power System Analysis

V1 @ Gen 1 = 1.00 pu − 0.353 pu @ − 87.06° × 0.370 pu @ 90°

V1 @ Gen 1 = 0.869 pu @ − 0.39°

V2 @ Gen 1 = −0.353 pu @ 92.61° × 0.370 pu @ 90°

V2 @ Gen 1 = 0.131 pu @ 2.61°

VA − N @ Gen 1 = 0.869 pu @ − 0.39° + 0.131 pu @ 2.61°

VA − N @ Gen 1 = 1.0 pu @ 0°

VA − N @ Gen 1 = 13.28 kV @ 0°

2.11 Zero Sequence Networks Zero sequence networks, like negative sequence networks, do not contain voltage sources because machines (generators or motors) only generate positive sequence voltages. However, the number and connection of branches will be strongly influenced by transformer winding connections. Most zero sequence networks consist of an array of separate zero sequence networks, as shown in Figure 2.14. Zero sequence line impedances are obtained by imagining three conductors connected together at the point of fault with the ground forming the return conductor. Checking transformer winding connections, and properly accounting for them in zero sequence networks, is a necessary requirement when calculating phase to ground fault current and voltage. The connection to the zero sequence reference bus changes the impedance of the zero sequence network and complicates circuit analysis as shown in Figure 2.15. Transformer equivalent circuits depend on the type of connection for each winding, delta or wye, and if wye, whether the transformer is solidly grounded, resistance grounded, reactance grounded, or ungrounded. Figure 2.16 illustrates transformer winding connections (delta, wye, and wye-grounded) Bus G1 RB0

Z0 Gen 1

Bus 1

RB0

Z0 Trans 1

Bus T2H Z0 Line

Bus T2L

RB0

Z0 Trans 2

FIGURE 2.14 Zero sequence network for power system shown in Figure 2.9.

Bus G1 RB0

Z0 Gen 1 0.012 + j0.474 pu

Bus 1

Bus T2H

Z0 Line Z Trans 1 RB0 0 0.005 + j0.135 pu 0.086 + j0.729 pu

Base MVA = 100 MVA Base voltages = 23, 69, and 13.8 kV

FIGURE 2.15 Zero sequence network with normalized impedances.

Bus T2L Z Trans 2 RB0 0 0.029 + j0.720 pu

2-14

Power Systems Phase 1

Positive/negative sequence equivalent circuit Trans. lead

Phase 2

Trans. midpoint

Zero sequence equivalent circuit

Phase 3

Trans. lead

Trans. midpoint

RB0

(a)

Phase 1 Positive/negative sequence equivalent circuit

Phase 2

Neutral

Trans. lead

Trans. midpoint

Trans. lead

Trans. midpoint

Phase 3 (b) Positive/negative/zero sequence equivalent circuit

Phase 1

Trans. lead

Phase 2

Trans. midpoint

Neutral

Phase 3 (c)

FIGURE 2.16 (a) Transformer windings connected phase to phase—“delta.” (b) Transformer windings connected phase to neutral—“wye.” (c) Transformer windings connected phase to ground—“wye-grounded.”

and the equivalent circuit for each sequence. When using symmetrical components, transformers are represented as “star equivalents.” High-voltage terminals and low-voltage terminals are located in the network and then connections from a midpoint to each terminal is established using the equivalents shown in Figure 2.16. When transformers are connected wye-grounded/wye-grounded, the equivalent circuit is the same in the positive, negative, and zero sequences, as shown in Figure 2.17. For two winding transformers, the impedance listed on the nameplate is the sum of the high-voltage winding impedance and the low voltage winding impedance. For autotransformers and other transformers with more than two windings, impedances must be converted to a “star equivalent.” In wye-grounded/wye-grounded transformers, I0, zero sequence current can flow in each winding, in each phase, and in each transformer lead. 3I0 can flow in the ground circuit. When transformers are connected delta/wye-grounded, the zero sequence equivalent circuit, shown in Figure 2.18, is different from the positive and negative sequence equivalent circuit. Note that the zero sequence equivalent circuit shows (1) an open circuit between the transformer leads connected to the delta winding and the transformer midpoint and (2) a connection from the transformer midpoint to the zero sequence reference bus.

2-15

Symmetrical Components for Power System Analysis H

X

Positive/negative/zero sequence equivalent circuit Trans. midpoint

Trans. lead H winding

Trans. lead X winding

FIGURE 2.17 Transformer equivalent circuit, windings connected wye grounded-wye grounded.

Positive/negative sequence equivalent circuit Trans. lead delta winding

Trans. midpoint

Trans. lead wye winding

Zero sequence equivalent circuit Trans. midpoint

Trans. lead delta winding

Trans. lead wye winding

RB0

FIGURE 2.18 Transformer equivalent circuit, windings connected delta-wye grounded.

In delta-wye grounded transformers, I0, can flow in each winding, in each phase, but not in the transformer leads connected to the delta winding. 3I0 can flow in the ground circuit connected to the wyegrounded winding. When transformers are connected delta/wye, wye/wye, or wye/wye-grounded, the equivalent circuit in the zero sequence is an open circuit. The equivalent circuit for a transformer connected delta/wye is shown in Figure 2.19. Zero sequence impedance of generators, like transformers, depends on the type of winding connection, delta or wye, and if wye, whether the generator is solidly grounded, resistance grounded, reactance grounded, or ungrounded. Adjacent overhead transmission lines influence the zero sequence impedance of each other. This influence is referred to as mutual impedance. Figure 2.13 is the negative sequence network with normalized impedances. To calculate a phase to phase fault, the positive and negative sequence networks must be connected as shown in Figure 2.3. This means that one jumper is needed between RB1 and RB2 and a second jumper is needed between the positive sequence fault bus and the corresponding negative sequence fault bus.

2-16

Power Systems

Positive/negative sequence equivalent circuit Trans. midpoint Trans. lead delta winding

Trans. lead wye winding

Zero sequence equivalent circuit Trans. midpoint

Trans. lead delta winding

Trans. lead wye winding

RB0

FIGURE 2.19 Transformer equivalent circuits, windings connected delta-wye.

2.12 Sample Phase to Ground Fault Calculation: Fault at Bus T2L

V = 1.00 pu = 2 × (0.064 + j1.414) + .029 + j0.720 pu Z = Z1 + Z 2 + Z 0 = 0.158 + j3.548 pu = 3.552 pu @ 87.45° Ia1 = 0.282 pu @ − 87.45° Ia2 = 0.282 pu @ − 87.45° Ia0 = 0.282 pu @ − 87.45° Ia1 = 1178 A @ − 87.45° at 13.8 kV Ia2 = 1178 A @ − 87.45° at 13.8 kV Ia0 = 1178 A @ − 87.45° at 13.8 kV IA = Ia1 + Ia2 + Ia 0 = 3534 A @ − 87.45° IB = a 2Ia1 + aIa2 + Ia 0 = 0

Symmetrical Components for Power System Analysis

IC = aIa1 + a 2Ia2 + Ia 0 = 0 V1 @ Gen 1 = 1.00 pu − 0.282 pu @ − 87.45° × 0.370 pu @ 90° V1 @ Gen 1 = 0.896 pu @ − 0.30° V2 @ Gen 1 = −0.282 pu @ − 87.45° × 0.370 pu @ 90° V2 @ Gen 1 = 0.104 pu @ − 2.55° V0 @ Gen 1 = 0.0 pu × 0.370 pu @ 90° V0 @ Gen 1 = 0.0 VA − N @ Gen 1 = 0.896 pu @ − 0.30° + 0.104 pu @ − 2.55° + 0 VA − N @ Gen 1 = 10.52 kV @ − 0.67°

2-17

3 Power Flow Analysis

Leonard L. Grigsby Auburn University

Andrew P. Hanson The Structure Group

3.1 3.2 3.3 3.4 3.5 3.6 3.7

Introduction....................................................................................... 3-1 Power Flow Problem.......................................................................... 3-1 Formulation of Bus Admittance Matrix........................................ 3-3 Formulation of Power Flow Equations........................................... 3-3 P–V Buses............................................................................................3-6 Bus Classifications............................................................................. 3-7 Generalized Power Flow Development.......................................... 3-7

3.8

Solution Methods...............................................................................3-8

Basic Power Flow Equations

Newton–Raphson Method • Fast Decoupled Power Flow Solution

3.9 Component Power Flows................................................................ 3-10 References..................................................................................................... 3-11 Further Information.................................................................................... 3-11

3.1 Introduction The equivalent circuit parameters of many power system components are described in other parts of this handbook. The interconnection of the different elements allows development of an overall power system model. The system model provides the basis for computational simulation of the system performance under a wide variety of projected operating conditions. Additionally, “post mortem” studies, performed after system disturbances or equipment failures, often provide valuable insight into contributing system conditions. This chapter discusses one such computational simulation, the power flow problem. Power systems typically operate under slowly changing conditions, which can be analyzed using steady-state analysis. Further, transmission systems operate under balanced or near-balanced conditions allowing per-phase analysis to be used with a high degree of confidence in the solution. Power flow analysis provides the starting point for most other analyses. For example, the small signal and transient stability effects of a given disturbance are dramatically affected by the “pre-disturbance” operating conditions of the power system. (A disturbance resulting in instability under heavily loaded system conditions may not have any adverse effects under lightly loaded conditions.) Additionally, fault analysis and transient analysis can also be impacted by the pre-disturbance operating point of the power system (although, they are usually affected much less than transient stability and small signal stability analysis).

3.2 Power Flow Problem Power flow analysis is fundamental to the study of power systems; in fact, power flow forms the core of power system analysis. A power flow study is valuable for many reasons. For example, power flow analyses play a key role in the planning of additions or expansions to transmission and generation facilities. A power flow solution is often the starting point for many other types of power system analyses. In addition, power flow analysis and many of its extensions are an essential ingredient of the studies 3-1

3-2

Power Systems

performed in power system operations. In this latter case, it is at the heart of contingency analysis and the implementation of real-time monitoring systems. The power flow problem (popularly known as the load flow problem) can be stated as follows: For a given power network, with known complex power loads and some set of specifications or restrictions on power generations and voltages, solve for any unknown bus voltages and unspecified generation and finally for the complex power flow in the network components. Additionally, the losses in individual components and the total network as a whole are usually calculated. Furthermore, the system is often checked for component overloads and voltages outside allowable tolerances. Balanced operation is assumed for most power flow studies and will be assumed in this chapter. Consequently, the positive sequence network is used for the analysis. In the solution of the power flow problem, the network element values are almost always taken to be in per-unit. Likewise, the calculations within the power flow analysis are typically in per-unit. However, the solution is usually expressed in a mixed format. Solution voltages are usually expressed in per-unit; powers are most often given in kVA or MVA. The “given network” may be in the form of a system map and accompanying data tables for the network components. More often, however, the network structure is given in the form of a one-line diagram (such as shown in Figure 3.1). Regardless of the form of the given network and how the network data is given, the steps to be followed in a power flow study can be summarized as follows:

1. Determine element values for passive network components. 2. Determine locations and values of all complex power loads. 3. Determine generation specifications and constraints. 4. Develop a mathematical model describing power flow in the network. 5. Solve for the voltage profile of the network. 6. Solve for the power flows and losses in the network. 7. Check for constraint violations. Steam

Hydro

10

9

6

7

5 2

1

8

Cogen

FIGURE 3.1 The one-line diagram of a power system.

4

3

3-3

Power Flow Analysis j +

Ij

Ye

c:1

Ik

k +

Vj

Vk

–

–

FIGURE 3.2 Off nominal turns ratio transformer.

3.3 Formulation of Bus Admittance Matrix The first step in developing the mathematical model describing the power flow in the network is the formulation of the bus admittance matrix. The bus admittance matrix is an n × n matrix (where n is the number of buses in the system) constructed from the admittances of the equivalent circuit elements of the segments making up the power system. Most system segments are represented by a combination of shunt elements (connected between a bus and the reference node) and series elements (connected between two system buses). Formulation of the bus admittance matrix follows two simple rules:

1. The admittance of elements connected between node k and reference is added to the (k, k) entry of the admittance matrix. 2. The admittance of elements connected between nodes j and k is added to the (j, j) and (k, k) entries of the admittance matrix. The negative of the admittance is added to the (j, k) and (k, j) entries of the admittance matrix.

Off nominal transformers (transformers with transformation ratios different from the system voltage bases at the terminals) present some special difficulties. Figure 3.2 shows a representation of an off nominal turns ratio transformer. The admittance matrix base mathematical model of an isolated off nominal transformer is

I j Y e − c Y e V j = * 2 I k − c Y e | c | Y e V k

(3.1)

where − Y e is the equivalent series admittance (referred to node j) − c is the complex (off nominal) turns ratio − I j is the current injected at node j − Vj is the voltage at node j (with respect to reference) Off nominal transformers are added to the bus admittance matrix by adding the corresponding entry of the isolated off nominal transformer admittance matrix to the system bus admittance matrix.

3.4 Formulation of Power Flow Equations Considerable insight into the power flow problem and its properties and characteristics can be obtained by consideration of a simple example before proceeding to a general formulation of the problem. This simple case will also serve to establish some notation.

3-4

Power Systems SG2

SG1

SD2

SD1 1

2

Network one-line diagram

SG3

3

4

SD3

SD4 SG4

FIGURE 3.3 Conceptual one-line diagram of a four-bus power system.

A conceptual representation of a one-line diagram for a four-bus power system is shown in Figure 3.3. For generality, we have shown a generator and a load connected to each bus. The following notation applies: − S GI = Complex power flow into bus 1 from the generator − S DI = Complex power flow into the load from bus 1 Comparable quantities for the complex power generations and loads are obvious for each of the three other buses. The positive sequence network for the power system represented by the one-line diagram of Figure 3.3 is shown in Figure 3.4. The boxes symbolize the combination of generation and load. Network texts refer to this network as a five-node network. (The balanced nature of the system allows analysis using only the positive sequence network; reducing each three-phase bus to a single node. The reference or ground represents the fifth node.) However, in power systems literature it is usually referred to as a fourbus network or power system. For the network of Figure 3.4, we define the following additional notation: − − − S = S G1 − S D1 = Net complex power injected at bus 1 −1 I 1 = Net positive sequence phasor current injected at bus 1 − V1 = Positive sequence phasor voltage at bus 1 The standard node voltage equations for the network can be written in terms of the quantities at bus 1 (defined above) and comparable quantities at the other buses:

I 1 = Y 11V 1 + Y 12V 2 + Y 13V 3 + Y 14V 4

(3.2)

I 2 = Y 21V 1 + Y 22V 2 + Y 23V 3 + Y 24V 4

(3.3)

I 3 = Y 31V 1 + Y 32V 2 + Y 33V 3 + Y 34V 4

(3.4)

I 4 = Y 41V 1 + Y 42V 2 + Y 43V 3 + Y 44V 4

(3.5)

3-5

Power Flow Analysis 2

1

I1

Network of passive components

S1

S2

+

+

3

V1 –

I2

I3

4 S3

I4

+

+

V3

V4

–

Neutral

S4

V2 –

–

FIGURE 3.4 Positive sequence network for the system in Figure 3.3.

− The admittances in Equations 3.2 through 3.5, Y ij, are the ijth entries of the bus admittance matrix for the power system. The unknown voltages could be found using linear algebra if the four currents − − I 1 … I 4 were known. However, these currents are not known. Rather, something is known about the complex power and voltage, at each bus. The complex power injected into bus k of the power system is defined by the relationship between complex power, voltage, and current given by the following equation:

S k = V k I *k

(3.6)

S k* = S *Gk − S *Dk V *k V *k

(3.7)

Therefore,

Ik=

By substituting this result into the nodal equations and rearranging, the basic power flow equations (PFE) for the four-bus system are given as follows:

* * S G1 − S D1 = V 1*[Y 11V 1 + Y 12V 2 + Y 13V 3 + Y 14V 4]

(3.8)

* S *G2 − S *D2 = V 2[Y 21V 1 + Y 22V 2 + Y 23V 3 + Y 24V 4]

(3.9)

* * * S G3 − S D3 = V 3[Y 31V 1 + Y 32V 2 + Y 33V 3 + Y 34V 4]

(3.10)

* * S G 4 − S D 4 = V *4[Y 41V 1 + Y 42V 2 + Y 43V 3 + Y 44V 4]

(3.11)

3-6

Power Systems

Examination of Equations 3.8 through 3.11 reveals that except for the trivial case where the generation equals the load at every bus, the complex power outputs of the generators cannot be arbitrarily selected. In fact, the complex power output of at least one of the generators must be calculated last since it must take up the unknown “slack” due to the, as yet, uncalculated network losses. Further, losses cannot be calculated until the voltages are known. These observations are the result of the principle of conservation of complex power (i.e., the sum of the injected complex powers at the four system buses is equal to the system complex power losses). Further examination of Equations 3.8 through 3.11 indicates that it is not possible to solve these equations for the absolute phase angles of the phasor voltages. This simply means that the problem can only be solved to some arbitrary phase angle reference. − In order to alleviate the dilemma outlined above, suppose S G4 is arbitrarily allowed to float or − − − swing (in order to take up the necessary slack caused by the losses) and that S G1, S G2 , and S G3 are specified (other cases will be considered shortly). Now, with the loads known, Equations 3.8 through 3.11 are seen as four simultaneous nonlinear equations with complex coefficients in five unknowns − − − − − V1, V2 , V3, V4, and S G4. The problem of too many unknowns (which would result in an infinite number of solutions) is solved by specifying another variable. Designating bus 4 as the slack bus and specifying the voltage − V4 reduces the problem to four equations in four unknowns. The slack bus is chosen as the phase reference for all phasor calculations, its magnitude is constrained, and the complex power generation at this bus is free to take up the slack necessary in order to account for the system real and reactive power losses. − The specification of the voltage V4 decouples Equation 3.11 from Equations 3.8 through 3.10, allowing calculation of the slack bus complex power after solving the remaining equations. (This property carries over to larger systems with any number of buses.) The example problem is reduced to solving only three − − − equations simultaneously for the unknowns V1, V2, and V3. Similarly, for the case of n buses it is necessary to solve n − 1 simultaneous, complex coefficient, nonlinear equations. Systems of nonlinear equations, such as Equations 3.8 through 3.10, cannot (except in rare cases) be solved by closed-form techniques. Direct simulation was used extensively for many years; however, essentially all power flow analyses today are performed using iterative techniques on digital computers.

3.5 P–V Buses In all realistic cases, the voltage magnitude is specified at generator buses to take advantage of the generator’s reactive power capability. Specifying the voltage magnitude at a generator bus requires a variable specified in the simple analysis discussed earlier to become an unknown (in order to bring the number of unknowns back into correspondence with the number of equations). Normally, the reactive power injected by the generator becomes a variable, leaving the real power and voltage magnitude as the specified quantities at the generator bus. It was noted earlier that Equation 3.11 is decoupled and only Equations 3.8 through 3.10 need be solved simultaneously. Although not immediately apparent, specifying the voltage magnitude at a bus and treating the bus reactive power injection as a variable result in retention of, effectively, the same number of complex unknowns. For example, if the voltage magnitude of bus 1 of the earlier four-bus system is specified and the reactive power injection at bus 1 becomes a variable, Equations 3.8 through − − 3.10 again effectively have three complex unknowns. (The phasor voltages V2 and V3 at buses 2 and 3 are two complex unknowns and the angle δ1 of the voltage at bus 1 plus the reactive power generation QG1 at bus 1 result in the equivalent of a third complex unknown.) Bus 1 is called a voltage controlled bus, since it is apparent that the reactive power generation at bus 1 is being used to control the voltage magnitude. This type of bus is also referred to as a P–V bus because of the specified quantities. Typically, all generator buses are treated as voltage controlled buses.

3-7

Power Flow Analysis

3.6 Bus Classifications There are four quantities of interest associated with each bus:

1. Real power, P 2. Reactive power, Q 3. Voltage magnitude, V 4. Voltage angle, δ

At every bus of the system two of these four quantities will be specified and the remaining two will be unknowns. Each of the system buses may be classified in accordance with which of the two quantities are specified. The following classifications are typical: Slack bus—The slack bus for the system is a single bus for which the voltage magnitude and angle are specified. The real and reactive power are unknowns. The bus selected as the slack bus must have a source of both real and reactive power, since the injected power at this bus must “swing” to take up the “slack” in the solution. The best choice for the slack bus (since, in most power systems, many buses have real and reactive power sources) requires experience with the particular system under study. The behavior of the solution is often influenced by the bus chosen. (In the earlier discussion, the last bus was selected as the slack bus for convenience.) Load bus (P–Q bus)—A load bus is defined as any bus of the system for which the real and reactive powers are specified. Load buses may contain generators with specified real and reactive power outputs; however, it is often convenient to designate any bus with specified injected complex power as a load bus. Voltage-controlled bus (P–V bus)—Any bus for which the voltage magnitude and the injected real power are specified is classified as a voltage controlled (or P–V) bus. The injected reactive power is a variable (with specified upper and lower bounds) in the power flow analysis. (A P–V bus must have a variable source of reactive power such as a generator or a capacitor bank.)

3.7 Generalized Power Flow Development The more general (n bus) case is developed by extending the results of the simple four-bus example. Consider the case of an n-bus system and the corresponding n + 1 node positive sequence network. Assume that the buses are numbered such that the slack bus is numbered last. Direct extension of the earlier equations (writing the node voltage equations and making the same substitutions as in the fourbus case) yields the basic power flow equations in the general form.

3.7.1 Basic Power Flow Equations Sk* = Pk − jQk = Vk*

n

∑Y V

for k = 1, 2, 3,…, n − 1

ki i

(3.12)

i =1

and

n

∑Y V

Pn − jQn = V n*

ni

i

(3.13)

i =1

Equation 3.13 is the equation for the slack bus. Equation 3.12 represents n − 1 simultaneous equations in n − 1 complex unknowns if all buses (other than the slack bus) are classified as load buses. Thus, given a set of specified loads, the problem is to solve Equation 3.12 for the n − 1 complex phasor voltages at the remaining buses. Once the bus voltages are known, Equation 3.13 can be used to calculate the slack bus power.

3-8

Power Systems

Bus j is normally treated as a P–V bus if it has a directly connected generator. The unknowns at bus j are then the reactive generation QGj and δj, because the voltage magnitude, Vj, and the real power generation, PGj, have been specified. The next step in the analysis is to solve Equation 3.12 for the bus voltages using some iterative method. Once the bus voltages have been found, the complex power flows and complex power losses in all of the network components are calculated.

3.8 Solution Methods The solution of the simultaneous nonlinear power flow equations requires the use of iterative techniques for even the simplest power systems. Although there are many methods for solving nonlinear equations, only two methods are discussed here.

3.8.1 Newton–Raphson Method The Newton–Raphson algorithm has been applied in the solution of nonlinear equations in many fields. The algorithm will be developed using a general set of two equations (for simplicity). The results are easily extended to an arbitrary number of equations. A set of two nonlinear equations are shown in the following equations:

f1(x1 , x2 ) = k1

(3.14)

f 2 (x1 , x2 ) = k2

(3.15)

Now, if x1(0) and x2(0) are inexact solution estimates and ∆x1(0) and ∆x2(0) are the corrections to the estimates to achieve an exact solution, Equations 3.14 and 3.15 can be rewritten as

f1(x1(0) + ∆x1(0) , x2(0) + ∆x2(0) ) = k1

(3.16)

f 2 (x1(0) + ∆x1(0) , x2(0) + ∆x2(0) ) = k2

(3.17)

Expanding Equations 3.16 and 3.17 in a Taylor series about the estimate yields:

(

)

∂f1 ∂x1

(

)

∂f 2 ∂x1

f1 x1(0) , x2(0) +

f 2 x1(0) , x2(0) +

(0 )

(0 )

∂f1 ∂x 2

(0 )

∆x1(0) +

∂f 2 ∂x 2

(0 )

∆x1(0) +

∆x2(0) + h.o.t. = k1

(3.18)

∆x2(0) + h.o.t. = k2

(3.19)

where the subscript, (0), on the partial derivatives indicates evaluation of the partial derivatives at the initial estimate and h.o.t. indicates the higher-order terms. Neglecting the higher-order terms (an acceptable approximation if ∆x1(0) and ∆x2(0) are small) Equations 3.18 and 3.19 can be rearranged and written in matrix form:

∂f1 ∂x1

(0 )

∂f 2 ∂x1

(0 )

(0 ) (0 ) 2 2

∂f1 ∂x 2 ∂f ∂x

(0 ) (0 ) ∆x1(0) k1 − f1(x1 , x2 ) ≈ ∆x2(0) k2 − f 2 (x1(0) , x2(0) )

(3.20)

3-9

Power Flow Analysis

The matrix of partial derivatives in Equation 3.20 is known as the Jacobian matrix and is evaluated at the initial estimate. Multiplying each side of Equation 3.20 by the inverse of the Jacobian matrix yields an approximation of the required correction to the estimated solution. Since the higher-order terms were neglected, addition of the correction terms to the original estimate will not yield an exact solution, but will often provide an improved estimate. The procedure may be repeated, obtaining successively better estimates until the estimated solution reaches a desired tolerance. Summarizing, correction terms for the ℓth iterate are given in Equation 3.21 and the solution estimate is updated according to Equation 3.22:

∂f1 ∂x1

()

∆x1( ) = () ∆x2( ) ∂f 2 ∂x1

() () 2 2

∂f1 ∂x 2

−1

∂f ∂x

k1 − f1(x1( ) , x2( ) ) () () k2 − f 2 (x1 , x2 )

x ( +1) = x ( ) + ∆x ( )

(3.21)

(3.22)

The solution of the original set of nonlinear equations has been converted to a repeated solution of a system of linear equations. This solution requires evaluation of the Jacobian matrix (at the current solution estimate) in each iteration. The power flow equations can be placed into the Newton–Raphson framework by separating the power flow equations into their real and imaginary parts and taking the voltage magnitudes and phase angles as the unknowns. Writing Equation 3.21 specifically for the power flow problem:

∂ P ( ) ∂P ( ) ∂V ∆ δ ( ) ∂δ () = ∆V ∂Q ( ) ∂Q ( ) ∂V ∂δ

−1

P (sched) − P ( ) () Q(sched) − Q

(3.23)

The underscored variables in Equation 3.23 indicate vectors (extending the two equation Newton– Raphson development to the general power flow case). The (sched) notation indicates the scheduled real and reactive powers injected into the system. P(ℓ) and Q(ℓ) represent the calculated real and reactive power injections based on the system model and the ℓth voltage phase angle and voltage magnitude estimates. The bus voltage phase angle and bus voltage magnitude estimates are updated, the Jacobian reevaluated, and the mismatch between the scheduled and calculated real and reactive powers evaluated in each iteration of the Newton–Raphson algorithm. Iterations are performed until the estimated solution reaches an acceptable tolerance or a maximum number of allowable iterations is exceeded. Once a solution (within an acceptable tolerance) is reached, P–V bus reactive power injections and the slack bus complex power injection may be evaluated.

3.8.2 Fast Decoupled Power Flow Solution The fast decoupled power flow algorithm simplifies the procedure presented for the Newton–Raphson algorithm by exploiting the strong coupling between real power and bus voltage phase angles and reactive power and bus voltage magnitudes commonly seen in power systems. The Jacobian matrix is simplified by approximating as zero the partial derivatives of the real power equations with respect to the bus voltage magnitudes. Similarly, the partial derivatives of the reactive power equations with respect to the

3-10

Power Systems

bus voltage phase angles are approximated as zero. Further, the remaining partial derivatives are often approximated using only the imaginary portion of the bus admittance matrix. These approximations yield the following correction equations: ∆δ ( ) = [B ′]−1[P (sched) − P ( )]

(3.24)

()

∆V ( ) = [B ′′]−1[Q(sched) − Q ]

(3.25)

where B′ is an approximation of the matrix of partial derivatives of the real power flow equations with respect to the bus voltage phase angles and B″ is an approximation of the matrix of partial derivatives of the reactive power flow equations with respect to the bus voltage magnitudes. B′ and B″ are typically held constant during the iterative process, eliminating the necessity of updating the Jacobian matrix (required in the Newton–Raphson solution) in each iteration. The fast decoupled algorithm has good convergence properties despite the many approximations used during its development. The fast decoupled power flow algorithm has found widespread use, since it is less computationally intensive (requires fewer computational operations) than the Newton–Raphson method.

3.9 Component Power Flows The positive sequence network for components of interest (connected between buses i and j) will be of the form shown in Figure 3.5. An admittance description is usually available from earlier construction of the nodal admittance matrix. Thus, I i Y a Y b V i = I j Y c Y d V j

(3.26)

Therefore the complex power flows and the component loss are

* Sij = V i I i* = V i [Y a V i + Y bV j]

(3.27)

S ji = V j I *j = V j [Y cV i + Y dV j ]*

(3.28)

Sloss = Sij + S ji

(3.29)

Ii

Sji

Sij

+

+ Vi –

FIGURE 3.5 Typical power system component.

Ij

Two port component

Vj –

Power Flow Analysis

3-11

The calculated component flows combined with the bus voltage magnitudes and phase angles provide extensive information about the power systems operating point. The bus voltage magnitudes may be checked to ensure operation within a prescribed range. The segment power flows can be examined to ensure no equipment ratings are exceeded. Additionally, the power flow solution may be used as the starting point for other analyses. An elementary discussion of the power flow problem and its solution is presented in this section. The power flow problem can be complicated by the addition of further constraints such as generator real and reactive power limits. However, discussion of such complications is beyond the scope of this chapter. The references provide detailed development of power flow formulation and solution under additional constraints. The references also provide some background in the other types of power system analysis.

References Bergen, A.R. and Vital, V., Power Systems Analysis, 2nd edn., Prentice-Hall, Inc., Englewood Cliffs, NJ, 2000. Elgerd, O.I., Electric Energy Systems Theory—An Introduction, 2nd edn., McGraw-Hill, New York, 1982. Glover, J.D. and Sarma, M., Power System Analysis and Design, 3rd edn., Brooks/Cole, Pacific Grove, CA, 2002. Grainger, J.J. and Stevenson, W.D., Elements of Power System Analysis, McGraw-Hill, New York, 1994. Gross, C.A., Power System Analysis, 2nd edn., John Wiley & Sons, New York, 1986.

Further Information The references provide clear introductions to the analysis of power systems. An excellent review of many issues involving the use of computers for power system analysis is provided in July 1974, Proceedings of the IEEE (Special Issue on Computers in the Power Industry). The quarterly journal IEEE Transactions on Power Systems provides excellent documentation of more recent research in power system analysis.

4 Fault Analysis in Power Systems

Charles A. Gross Auburn University

4.1 4.2

Simplifications in the System Model.............................................. 4-2 The Four Basic Fault Types...............................................................4-4

4.3

An Example Fault Study................................................................... 4-7

The Balanced Three-Phase Fault • The Single Phase-toGround Fault • The Phase-to-Phase Fault • The Double Phase-to-Ground Fault

Balanced Three-Phase Fault • Single Phase-to-Ground Fault

4.4 Further Considerations................................................................... 4-15 4.5 Summary........................................................................................... 4-16 4.6 Defining Terms................................................................................ 4-17 References..................................................................................................... 4-17 Further Information.................................................................................... 4-18

A fault in an electrical power system is the unintentional and undesirable creation of a conducting path (a short circuit) or a blockage of current (an open circuit). The short-circuit fault is typically the most common and is usually implied when most people use the term fault. We restrict our comments to the short-circuit fault. The causes of faults include lightning, wind damage, trees falling across lines, vehicles colliding with towers or poles, birds shorting out lines, aircraft colliding with lines, vandalism, small animals entering switchgear, and line breaks due to excessive ice loading. Power system faults may be categorized as one of four types: single line-to-ground, line-to-line, double line-to-ground, and balanced three-phase. The first three types constitute severe unbalanced operating conditions. It is important to determine the values of system voltages and currents during faulted conditions so that protective devices may be set to detect and minimize their harmful effects. The time constants of the associated transients are such that sinusoidal steady-state methods may still be used. The method of symmetrical components is particularly suited to fault analysis. Our objective is to understand how symmetrical components may be applied specifically to the four general fault types mentioned and how the method can be extended to any unbalanced three-phase system problem. Note that phase values are indicated by subscripts, a, b, c; sequence (symmetrical component) values are indicated by subscripts 0, 1, 2. The transformation is defined by

V 1 a V b = 1 V 1

1 a

2

a

1 V 0 V 0 a V 1 = [T ] V 1 V a2 V 2 2 4-1

4-2

Power Systems

4.1 Simplifications in the System Model Certain simplifications are possible and usually employed in fault analysis. • Transformer magnetizing current and core loss will be neglected. • Line shunt capacitance is neglected. • Sinusoidal steady-state circuit analysis techniques are used. The so-called DC offset is accounted for by using correction factors. • Prefault voltage is assumed to be 1∠0° per-unit. One per-unit voltage is at its nominal value prior to the application of a fault, which is reasonable. The selection of zero phase is arbitrary and convenient. Prefault load current is neglected. For hand calculations, neglect series resistance is usually neglected (this approximation will not be necessary for a computer solution). Also, the only difference in the positive and negative sequence networks is introduced by the machine impedances. If we select the subtransient reactance Xd′′ for the positive sequence reactance, the difference is slight (in fact, the two are identical for nonsalient machines). The simplification is important, since it reduces computer storage requirements by roughly one-third. Circuit models for generators, lines, and transformers are shown in Figures 4.1 through 4.3, respectively. Our basic approach to the problem is to consider the general situation suggested in Figure 4.4a. The general terminals brought out are for purposes of external connections that will simulate faults. Note carefully the positive assignments of phase quantities. Particularly note that the currents flow out of the system. We can construct general sequence equivalent circuits for the system, and such circuits are Excitation power components

Steam turbine Generator Controlled power rectifier assembly

Collector housing

Bus k Zn

X0

I0 +

3Zn

V0 –

k

X1 = Xd˝

(=X2)

+

~

–

Zero

FIGURE 4.1 Generator sequence circuit models.

I1 +

1 0° Positive

V1 –

k

X2

I2 + V2 –

Negative

k

4-3

Fault Analysis in Power Systems

Bus i

Bus j Single-line symbol

X1 = X2

i

j

Transmission line X0

i

Positive/negative sequence network

j

Zero sequence network

FIGURE 4.2 Line sequence circuit models.

2 1 3 Single-line symbol

1

X1(1)

X1(2)

2

X1(3)

3

Positive/negative sequence network The “gaps” are connected to model Y–∆ connections i i΄ 3Zn

1

1΄

X0(1)

X0(2)

2΄

2

X0(3)

3΄

3

Zero sequence network

FIGURE 4.3 Transformer sequence circuit models.

Zn i

i΄

4-4

Power Systems Ic

c

Ib

Power system phase (abc) network

+

b

Ia

a + Va –

+

Vc

Vb

n –

–

(a) I1

I0 + 0

I2 +

+ 1

V0 –

V1 –

V2 –

Negative

Positive

Zero

2

(b)

FIGURE 4.4 General fault port in an electric power system. (a) General fault port in phase (abc) coordinates; (b) corresponding fault ports in sequence (012) coordinates.

indicated in Figure 4.4b. The ports indicated correspond to the general three-phase entry port of Figure 4.4a. The positive sense of sequence values is compatible with that used for phase values.

4.2 The Four Basic Fault Types 4.2.1 The Balanced Three-Phase Fault Imagine the general three-phase access port terminated in a fault impedance (Z f ) as shown in Figure 4.5a. The terminal conditions are V a Z f V b = 0 V 0 c

0 Zf 0

0 Ia 0 I b Z f I c

Transforming to [Z012], Z f [Z012 ] = [T ] 0 0 −1

0 Zf 0

Z f 0 0 [T ] = 0 0 Zf

0 Zf 0

0 0 Zf

The corresponding network connections are given in Figure 4.6a. Since the zero and negative sequence networks are passive, only the positive sequence network is nontrivial.

V0 = V2 = 0

(4.1)

I0 = I2 = 0

(4.2)

V 1 = Z f I1

(4.3)

4-5

Fault Analysis in Power Systems a b c

a b c Ia + Va Zf

Ic

Ib + Vb Zf

Ib

Ia + Zf

+ Vc

Ic

+

Va

Zf

+

Vb

Vc

n

n (a)

(b) a b c

a b c Ia +

+

Va

Vb

Ia

Ic

Ib Zf

+ Vc

n

Ib

+

+

Va

Vb

n

Ic + Zf

Vc

(d)

(c)

FIGURE 4.5 Fault types. (a) Three-phase fault; (b) single phase-to-ground fault; (c) phase-to-phase fault; (d) double phase-to-ground fault.

4.2.2 The Single Phase-to-Ground Fault Imagine the general three-phase access port terminated as shown in Figure 4.5b. The terminal conditions are

I b = 0 I c = 0 V a = I aZ f

Therefore,

2

I 0 + a I 1 + aI 2 = I 0 + aI 1 + a 2 I 2 = 0

or I1 = I 2

Also,

2

I b = I 0 + a I 1 + aI 2 = I 0 + (a2 + a)I 1 = 0

or

I 0 = I1 = I 2

(4.4)

Furthermore, it is required that V a = Z f Ia

V 0 + V 1 + V 2 = 3Z f I 1

(4.5)

4-6

Power Systems I0

I0 +

Zero sequence network

V0 – I1 V1 –

+

Positive sequence network

Zf

3Zf

V1 – I2

I2 +

Negative sequence network

V0 – I1

+

Positive sequence network

+

Zero sequence network

Zf

V2 –

+

Negative sequence network

Zf

(a)

V2 –

(b) I0

I0 +

Zero sequence network

V0 – I1 +

Positive sequence network

V1 –

Zf

3Zf

V0 – I1 +

Positive sequence network

I2 Negative sequence network

+

Zero sequence network

V1 – I2

+ V2 –

(c)

Negative sequence network

+ V2 –

(d)

FIGURE 4.6 Sequence network terminations for fault types. (a) Balanced three-phase fault; (b) single phase-toground fault; (c) phase-to-phase fault; (d) double phase-to-ground fault.

In general then, Equations 4.4 and 4.5 must be simultaneously satisfied. These conditions can be met by interconnecting the sequence networks as shown in Figure 4.6b.

4.2.3 The Phase-to-Phase Fault Imagine the general three-phase access port terminated as shown in Figure 4.5c. The terminal conditions are such that we may write

I 0 = 0 I b = −I c V b = Z f I b + V c

It follows that

I 0 + I1 + I2 = 0

(4.6)

4-7

Fault Analysis in Power Systems

I0 = 0

(4.7)

I1 = −I 2

(4.8)

In general then, Equations 4.6 through 4.8 must be simultaneously satisfied. The proper interconnection between sequence networks appears in Figure 4.6c.

4.2.4 The Double Phase-to-Ground Fault Consider the general three-phase access port terminated as shown in Figure 4.5d. The terminal conditions indicate I a = 0 Vb = Vc Vb = ( Ib + I c )Z f

It follows that

I 0 + I1 + I2 = 0

(4.9)

V1 = V 2

(4.10)

V 0 − V 1 = 3Z f I 0

(4.11)

and

For the general double phase-to-ground fault, Equations 4.9 through 4.11 must be simultaneously satisfied. The sequence network interconnections appear in Figure 4.6d.

4.3 An Example Fault Study Case: EXAMPLE SYSTEM Run: System has data for two Line(s); two Transformer(s); four Bus(es); and two Generator(s) Transmission line data Bus

Bus

Seq

R

X

B

Srat

1

2

3

2

3

0.00000 0.00000 0.00000 0.00000

0.16000 0.50000 0.16000 0.50000

0.00000 0.00000 0.00000 0.00000

1.0000

2

pos zero pos zero

Line

1.0000

4-8

Power Systems

Transformer data Transformer

HV Bus

LV Bus

Seq

R

X

C

Srat

2 Y 3 Y

1 Y 4 D

pos zero pos zero

0.00000 0.00000 0.00000 0.00000

0.05000 0.05000 0.05000 0.05000

1.00000

1.0000

1.00000

1.0000

1 2

Generator data No. 1 2

Bus

Srated

Ra

Xd′′

Xo

Rn

Xn

Con

1 4

1.0000 1.0000

0.0000 0.0000

0.200 0.200

0.0500 0.0500

0.0000 0.0000

0.0400 0.0400

Y Y

Zero sequence {Z} matrix 0.0 + j(0.1144) 0.0 + j(0.0981) 0.0 + j(0.0163) 0.0 + j(0.0000)

0.0 + j(0.0981) 0.0 + j(0.1269) 0.0 + j(0.0212) 0.0 + j(0.0000)

0.0 + j(0.0163) 0.0 + j(0.0212) 0.0 + j(0.0452) 0.0 + j(0.0000)

0.0 + j(0.1138) 0.0 + j(0.1422) 0.0 + j(0.1078) 0.0 + j(0.0862)

0.0 + j(0.0862) 0.0 + j(0.1078) 0.0 + j(0.1422) 0.0 + j(0.1138)

0.0 + j(0.0000) 0.0 + j(0.0000) 0.0 + j(0.0000) 0.0 + j(0.1700)

Positive sequence [Z] matrix 0.0 + j(0.1310) 0.0 + j(0.1138) 0.0 + j(0.0862) 0.0 + j(0.0690)

0.0 + j(0.0690) 0.0 + j(0.0862) 0.0 + j(0.1138) 0.0 + j(0.1310)

The single-line diagram and sequence networks are presented in Figure 4.7. Suppose bus 3 in the example system represents the fault location and Z f = 0. The positive sequence circuit can be reduced to its Thévenin equivalent at bus 3: ET 1 = 1.0∠0 Z T 1 = j0.1422

1

2

3

0.16 4

0.04

0.04

(a)

+

0.20 1 0.05 2

0.12 (b)

0.50

3 0.05

4 0.20

– (c)

+ –

0.16

0.50 0.05 1 0.05 2

0.16

3 0.05

4 0.05

0.20 1 0.05 2

0.16

3 0.05

4 0.20

0.12 (d)

FIGURE 4.7 Example system. (a) Single-line diagram; (b) zero sequence network; (c) positive sequence network; (d) negative sequence network.

4-9

Fault Analysis in Power Systems I0

j.0452 Zero sequence

1 0°

+

~

–

j.0452

+

Positive sequence

+

1 0°

V1 –

~

–

j0.1422

V0 – I1 +

Positive sequence j0.1422

+

Negative sequence

V0 – I2 +

Negative sequence

V2 –

(a)

V2 –

(b) I0

j.0452

+

~

–

j0.1422

I1 +

1 0°

V1 –

+

~

–

I2

Negative sequence

I0 +

Zero sequence

V0 –

Positive sequence j0.1422

j.0452

+

Zero sequence

1 0°

+

I2

j0.1422

+

Zero sequence

V0 – I1

j0.1422

I0

j0.1422

I1

Positive sequence j0.1422

+

V0 –

(c)

V1 –

I2

Negative sequence

V2 –

+

+ V2 –

(d)

FIGURE 4.8 Example system faults at bus 3. (a) Balanced three-phase; (b) single phase-to-ground; (c) phase-tophase; (d) double phase-to-ground.

Similarly, the negative and zero sequence Thévenin elements are ET 2 = 0 Z T 2 = j0.1422 ET 0 = 0 ZT 0 = j0.0452

The network interconnections for the four fault types are shown in Figure 4.8. For each of the fault types, compute the currents and voltages at the faulted bus.

4.3.1 Balanced Three-Phase Fault The sequence networks are shown in Figure 4.8a. Obviously, V 0 = I0 = V 2 = I2 = 0

I1 =

1∠0 = − j7.032; also V 1 = 0 j0.1422

4-10

Power Systems

To compute the phase values, I a I 0 1 I b = [T ] I 1 = 1 I I 1 c 2

1 0 7.032∠− 90 a − j7.032 = 7.032∠150 a2 0 7.032∠30

1 a2 a

V a 0 0 V b = [T ] 0 = 0 V c 0 0

4.3.2 Single Phase-to-Ground Fault The sequence networks are interconnected as shown in Figure 4.8b. 1∠0 = − j3.034 j0.0452 + j0.1422 + j0.1422

I 0 = I1 = I 2 = I a 1 I b = 1 I 1 c

1 a2 a

1 − j3.034 − j9.102 a − j3.034 = 0 a2 − j3.034 0

The sequence voltages are V 0 = − j0.0452(− j3.034) = −1371 V 1 = 1.0 − j0.1422(− j3.034) = 0.5685 V 2 = − j0.1422(− j3.034) = −0.4314

The phase voltages are

V a 1 V b = 1 V 1 c

1 a2 a

1 −0.1371 0 a 0.5685 = 0.8901∠−103.4 a2 −0.4314 0.8901∠−103.4

Phase-to-phase and double phase-to-ground fault values are calculated from the appropriate networks (Figure 4.8c and d). Complete results are provided. Faulted Bus 3

Phase a

Phase b

Phase c

G

G

G

Sequence voltages Bus 1 2 3 4

V0 0.0000/ 0.0000/ 0.0000/ 0.0000/

V1 0.0 0.0 0.0 0.0

0.3939/ 0.2424/ 0.0000/ 0.2000/

V2 0.0 0.0 0.0 −30.0

0.0000/ 0.0000/ 0.0000/ 0.0000/

0.0 0.0 0.0 30.0

4-11

Fault Analysis in Power Systems

Phase voltages Va

Bus 1 2 3 4

Vb

0.3939/ 0.2424/ 0.0000/ 0.2000/

0.0 0.0 6.5 −30.0

0.3939/ 0.2424/ 0.0000/ 0.2000/

Vc −120.0 −120.0 −151.2 −150.0

0.3939/ 0.2424/ 0.0000/ 0.2000/

120.0 120.0 133.8 90.0

Sequence currents

I0

Bus to Bus 1 1 2 2 2 3 3 3 4 4

2 0 3 3 1 2 2 4 3 0

0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/

I1 167.8 −12.2 167.8 167.8 −12.2 −12.2 −12.2 −12.2 0.0 0.0

Faulted Bus

3.0303/ 3.0303/ 1.5152/ 1.5152/ 3.0303/ 1.5152/ 1.5152/ 4.0000/ 4.0000/ 4.0000/

I2 −90.0 90.0 −90.0 −90.0 90.0 90.0 90.0 90.0 −120.0 60.0

0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/

Phase a

Phase b

Phase c

G

G

G

3

90.0 −90.0 90.0 90.0 −90.0 −90.0 −90.0 −90.0 120.0 −60.0

Phase currents

Ia

Bus to Bus 1 1 2 2 2 3 3 3 4 4

2 0 3 3 1 2 2 4 3 0

3.0303/ 3.0303/ 1.5151/ 1.5151/ 3.0303/ 1.5151/ 1.5151/ 4.0000/ 4.0000/ 4.0000/

Ib −90.0 90.0 −90.0 −90.0 90.0 90.0 90.0 90.0 −120.0 60.0

Faulted Bus 3

3.0303/ 3.0303/ 1.5151/ 1.5151/ 3.0303/ 1.5151/ 1.5151/ 4.0000/ 4.0000/ 4.0000/

Ic 150.0 −30.0 150.0 150.0 −30.0 −30.0 −30.0 −30.0 120.0 −60.0

3.0303/ 3.0303/ 1.5151/ 1.5151/ 3.0303/ 1.5151/ 1.5151/ 4.0000/ 4.0000/ 4.0000/

Phase a

Phase b

Phase c

G

0

0

30.0 −150.0 30.0 30.0 −150.0 −150.0 −150.0 −150.0 −0.0 −180.0

4-12

Power Systems

Sequence voltages V0

Bus 1 2 3 4

V1

0.0496/ 0.0642/ 0.1371/ 0.0000/

180.0 180.0 180.0 0.0

0.7385/ 0.6731/ 0.5685/ 0.6548/

V2 0.0 0.0 0.0 −30.0

0.2615/ 0.3269/ 0.4315/ 0.3452/

180.0 180.0 180.0 210.0

Phase voltages Va

Bus 1 2 3 4

0.4274/ 0.2821/ 0.0000/ 0.5674/

Vb 0.0 0.0 89.2 −61.8

0.9127/ 0.8979/ 0.8901/ 0.5674/

Vc −108.4 −105.3 −103.4 −118.2

0.9127/ 0.8979/ 0.8901/ 1.0000/

108.4 105.3 103.4 90.0

Sequence currents I0

Bus to Bus 1 1 2 2 2 3 3 3 4 4

2 0 3 3 1 2 2 4 3 0

0.2917/ 0.2917/ 0.1458/ 0.1458/ 0.2917/ 0.1458/ 0.1458/ 2.7416/ 0.0000/ 0.0000/

Faulted Bus

I1 −90.0 90.0 −90.0 −90.0 90.0 90.0 90.0 90.0 0.0 90.0

1.3075/ 1.3075/ 0.6537/ 0.6537/ 1.3075/ 0.6537/ 0.6537/ 1.7258/ 1.7258/ 1.7258/

I2 −90.0 90.0 −90.0 −90.0 90.0 90.0 90.0 90.0 −120.0 60.0

1.3075/ 1.3075/ 0.6537/ 0.6537/ 1.3075/ 0.6537/ 0.6537/ 1.7258/ 1.7258/ 1.7258/

Phase a

Phase b

Phase c

G

0

0

3

−90.0 90.0 −90.0 −90.0 90.0 90.0 90.0 90.0 −60.0 120.0

Phase currents Ia

Bus to Bus 1 1 2 2 2 3 3 3 4 4

2 0 3 3 1 2 2 4 3 0

2.9066/ 2.9066/ 1.4533/ 1.4533/ 2.9066/ 1.4533/ 1.4533/ 6.1933/ 2.9892/ 2.9892/

Ib −90.0 90.0 −90.0 −90.0 90.0 90.0 90.0 90.0 −90.0 90.0

1.0158/ 1.0158/ 0.5079/ 0.5079/ 1.0158/ 0.5079/ 0.5079/ 1.0158/ 2.9892/ 2.9892/

Ic 90.0 −90.0 90.0 90.0 −90.0 −90.0 −90.0 90.0 90.0 −90.0

1.0158/ 1.0158/ 0.5079/ 0.5079/ 1.0158/ 0.5079/ 0.5079/ 1.0158/ 0.0000/ 0.0000/

90.0 −90.0 90.0 90.0 −90.0 −90.0 −90.0 90.0 −90.0 90.0

4-13

Fault Analysis in Power Systems

Faulted Bus

Phase a

Phase b

Phase c

0

C

B

3

Sequence voltages V0

Bus 1 2 3 4

V1

0.0000/ 0.0000/ 0.0000/ 0.0000/

0.0 0.0 0.0 0.0

V2

0.6970/ 0.6212/ 0.5000/ 0.6000/

0.0 0.0 0.0 −30.0

0.3030/ 0.3788/ 0.5000/ 0.4000/

0.0 0.0 0.0 30.0

Phase voltages Va

Bus 1 2 3 4

Vb

1.0000/ 1.0000/ 1.0000/ 0.8718/

0.0 0.0 0.0 −6.6

Vc

0.6053/ 0.5423/ 0.5000/ 0.8718/

−145.7 −157.2 −180.0 −173.4

0.6053/ 0.5423/ 0.5000/ 0.2000/

145.7 157.2 −180.0 90.0

Sequence currents I0

Bus to Bus 1 1 2 2 2 3 3 3 4 4

2 0 3 3 1 2 2 4 3 0

0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ Faulted Bus

I1 −61.0 119.0 −61.0 −61.0 119.0 119.0 119.0 119.0 0.0 90.0

1.5152/ 1.5152/ 0.7576/ 0.7576/ 1.5152/ 0.7576/ 0.7576/ 2.0000/ 2.0000/ 2.0000/

I2 −90.0 90.0 −90.0 −90.0 90.0 90.0 90.0 90.0 −120.0 60.0

1.5152/ 1.5152/ 0.7576/ 0.7576/ 1.5152/ 0.7576/ 0.7576/ 2.0000/ 2.0000/ 2.0000/

Phase a

Phase b

Phase c

0

C

B

3

90.0 −90.0 90.0 90.0 −90.0 −90.0 −90.0 −90.0 120.0 −60.0

Phase currents Ia

Bus to Bus 1 1 2 2 2 3 3 3 4 4

2 0 3 3 1 2 2 4 3 0

0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 0.0000/ 2.0000/ 2.0000/

Ib 180.0 180.0 −180.0 −180.0 180.0 −180.0 −180.0 −180.0 −180.0 0.0

2.6243/ 2.6243/ 1.3122/ 1.3122/ 2.6243/ 1.3122/ 1.3122/ 3.4641/ 2.0000/ 2.0000/

Ic 180.0 0.0 180.0 180.0 0.0 0.0 0.0 0.0 180.0 0.0

2.6243/ 2.6243/ 1.3122/ 1.3122/ 2.6243/ 1.3122/ 1.3122/ 3.4641/ 4.0000/ 4.0000/

0.0 180.0 0.0 0.0 180.0 180.0 180.0 180.0 0.0 −180.0

4-14

Power Systems

Faulted Bus

Phase a

Phase b

Phase c

0

G

G

3

Sequence voltages V0

Bus 1 2 3 4

V1

0.0703/ 0.0909/ 0.1943/ 0.0000/

0.0 0.0 −0.0 0.0

0.5117/ 0.3896/ 0.1943/ 0.3554/

0.0 0.0 0.0 −12.7

0.4197/ 0.2749/ 0.0000/ 0.4536/

V2 0.0 0.0 0.0 −30.0

0.1177/ 0.1472/ 0.1943/ 0.1554/

0.0 0.0 0.0 30.0

Phase voltages Va

Bus 1 2 3 4

Vb

0.6997/ 0.6277/ 0.5828/ 0.4536/

Vc −125.6 −130.2 −30.7 −167.3

0.4197/ 0.2749/ 0.0000/ 0.2000/

125.6 130.2 −139.6 90.0

Sequence currents I0

Bus to Bus 1 1 2 2 2 3 3 3 4 4

2 0 3 3 1 2 2 4 3 0

0.4133/ 0.4133/ 0.2067/ 0.2067/ 0.4133/ 0.2067/ 0.2067/ 3.8854/ 0.0000/ 0.0000/

I1 90.0 −90.0 90.0 90.0 −90.0 −90.0 −90.0 −90.0 0.0 −90.0

Faulted Bus 3

I2

2.4416/ 2.4416/ 1.2208/ 1.2208/ 2.4416/ 1.2208/ 1.2208/ 3.2229/ 3.2229/ 3.2229/

−90.0 90.0 −90.0 −90.0 90.0 90.0 90.0 90.0 −120.0 60.0

0.5887/ 0.5887/ 0.2943/ 0.2943/ 0.5887/ 0.2943/ 0.2943/ 0.7771/ 0.7771/ 0.7771/

Phase a

Phase b

Phase c

0

G

G

90.0 −90.0 90.0 90.0 −90.0 −90.0 −90.0 −90.0 120.0 −60.0

Phase currents Ia

Bus to Bus 1 1 2 2 2 3 3 3 4 4

2 0 3 3 1 2 2 4 3 0

1.4396/ 1.4396/ 0.7198/ 0.7198/ 1.4396/ 0.7198/ 0.7198/ 1.4396/ 2.9132/ 2.9132/

Ib −90.0 90.0 −90.0 −90.0 90.0 90.0 90.0 −90.0 −133.4 46.6

2.9465/ 2.9465/ 1.4733/ 1.4733/ 2.9465/ 1.4733/ 1.4733/ 6.1721/ 2.9132/ 2.9132/

Ic 153.0 −27.0 153.0 153.0 −27.0 −27.0 −27.0 −55.9 133.4 −46.6

2.9465/ 2.9465/ 1.4733/ 1.4733/ 2.9465/ 1.4733/ 1.4733/ 6.1721/ 4.0000/ 4.0000/

27.0 −153.0 27.0 27.0 −153.0 −153.0 −153.0 −124.1 −0.0 −180.0

4-15

Fault Analysis in Power Systems

4.4 Further Considerations Generators are not the only sources in the system. All rotating machines are capable of contributing to fault current, at least momentarily. Synchronous and induction motors will continue to rotate due to inertia and function as sources of fault current. The impedance used for such machines is usually the transient reactance Xd′ or the subtransient Xd′′ , depending on protective equipment and speed of response. Frequently, motors smaller than 50 hp are neglected. Connecting systems are modeled with their Thévenin equivalents. Although we have used AC circuit techniques to calculate faults, the problem is fundamentally transient since it involves sudden switching actions. Consider the so-called DC offset current. We model the system by determining its positive sequence Thévenin equivalent circuit, looking back into the positive sequence network at the fault, as shown in Figure 4.9. The transient fault current is i(t ) = I AC 2 cos(ωt − β) + I DCe −t / τ

This is a first-order approximation and strictly applies only to the three-phase or phase-to-phase fault. Ground faults would involve the zero sequence network also. I AC =

E 2

R + X2

= rms AC current

I DC (t ) = I DCe −t / τ = DC offset current

The maximum initial DC offset possible would be Max I DC = I max = 2 I AC

The DC offset will exponentially decay with time constant τ, where τ=

L X = R ωR

The maximum DC offset current would be IDC(t) I DC (t ) = I DCe −t/τ = 2 I ACe −t/τ

The transient rms current I(t), accounting for both the AC and DC terms, would be 2 2 I (t ) = I AC + I DC (t ) = I AC 1 + 2e −2t/τ

jx

R

+

~

E –

FIGURE 4.9 Positive sequence circuit looking back into faulted bus.

4-16

Power Systems

Define a multiplying factor ki such that IAC is to be multiplied by ki to estimate the interrupting capacity of a breaker which operates in time Top. Therefore, ki =

I (Top ) −2T / τ = 1 + 2e op I AC

Observe that the maximum possible value for ki is √3. Example 4.1 In the circuit of Figure 4.9, E = 2400 V, X = 2 Ω, R = 0.1 Ω, and f = 60 Hz. Compute ki and determine the interrupting capacity for the circuit breaker if it is designed to operate in two cycles. The fault is applied at t = 0. Solution I AC ≅

2400 = 1200 A 2

Top =

2 = 0.0333 s 60

τ=

2 X = = 0.053 ωR 37.7

ki = 1 + 2e

−2Top /τ

= 1 + 2e −0.0067/ 0.053 = 1.252

Therefore, I = ki I AC = 1.252(1200) = 1503 A

The Thévenin equivalent at the fault point is determined by normal sinusoidal steady-state methods, resulting in a first-order circuit as shown in Figure 4.9. While this provides satisfactory results for the steady-state component IAC , the X/R value so obtained can be in serious error when compared with the rate of decay of I(t) as measured by oscillographs on an actual faulted system. The major reasons for the discrepancy are, first of all, that the system, for transient analysis purposes, is actually high-order, and second, the generators do not hold constant impedance as the transient decays.

4.5 Summary Computation of fault currents in power systems is best done by computer. The major steps are summarized below: • • • • • • •

Collect, read in, and store machine, transformer, and line data in per-unit on common bases. Formulate the sequence impedance matrices. Define the faulted bus and Zf. Specify type of fault to be analyzed. Compute the sequence voltages. Compute the sequence currents. Correct for wye-delta connections. Transform to phase currents and voltages.

For large systems, computer formulation of the sequence impedance matrices is required. Refer to Further Information for more detail. Zero sequence networks for lines in close proximity to each other

4-17

Fault Analysis in Power Systems

(on a common right-of-way) will be mutually coupled. If we are willing to use the same values for positive and negative sequence machine impedances, [Z1] = [Z2 ]

Therefore, it is unnecessary to store these values in separate arrays, simplifying the program and reducing the computer storage requirements significantly. The error introduced by this approximation is usually not important. The methods previously discussed neglect the prefault, or load, component of current; that is, the usual assumption is that currents throughout the system were zero prior to the fault. This is almost never strictly true; however, the error produced is small since the fault currents are generally much larger than the load currents. Also, the load currents and fault currents are out of phase with each other, making their sum more nearly equal to the larger components than would have been the case if the currents were in phase. In addition, selection of precise values for prefault currents is somewhat speculative, since there is no way of predicting what the loaded state of the system is when a fault occurs. When it is important to consider load currents, a power flow study is made to calculate currents throughout the system, and these values are superimposed on (added to) results from the fault study. A term which has wide industrial use and acceptance is the fault level or fault MVA at a bus. It relates to the amount of current that can be expected to flow out of a bus into a three-phase fault. As such, it is an alternate way of providing positive sequence impedance information. Define Fault level in MVA at bus i = Vipunominal Iipufault S3φbase

S3φbase = (1) 11 S3φ base = 1 Zii Zii

Fault study results may be further refined by approximating the effect of DC offset. The basic reason for making fault studies is to provide data that can be used to size and set protective devices. The role of such protective devices is to detect and remove faults to prevent or minimize damage to the power system.

4.6 Defining Terms DC offset—The natural response component of the transient fault current, usually approximated with a first-order exponential expression. Fault—An unintentional and undesirable conducting path in an electrical power system. Fault MVA—At a specific location in a system, the initial symmetrical fault current multiplied by the prefault nominal line-to-neutral voltage (×3 for a three-phase system). Sequence (012) quantities—Symmetrical components computed from phase (abc) quantities. Can be voltages, currents, and/or impedances.

References Anderson P.M., Analysis of Faulted Power Systems, Ames, IA: Iowa State Press, 1973. Elgerd O.I., Electric Energy Systems Theory: An Introduction, 2nd edn., New York: McGraw-Hill, 1982. El-Hawary M.E., Electric Power Systems: Design and Analysis, Reston, VA: Reston Publishing, 1983. El-Hawary M.E., Electric Power Systems, New York: IEEE Press, 1995. General Electric, Short-circuit current calculations for industrial and commercial power systems, Publication GET-3550.

4-18

Power Systems

Gross C.A., Power System Analysis, 2nd edn., New York: Wiley, 1986. Horowitz S.H., Power System Relaying, 2nd edn., New York: Wiley, 1995. Lazar I., Electrical Systems Analysis and Design for Industrial Plants, New York: McGraw-Hill, 1980. Mason C.R., The Art and Science of Protective Relaying, New York: Wiley, 1956. Neuenswander J.R., Modern Power Systems, Scranton, PA: International Textbook, 1971. Stagg G. and El-Abiad A.H., Computer Methods in Power System Analysis, New York: McGraw-Hill, 1968. Westinghouse Electric Corporation, Applied Protective Relaying, Newark, NJ: Relay-Instrument Division, 1976. Wood A.J., Power Generation, Operation, and Control, New York: Wiley, 1996.

Further Information For a comprehensive coverage of general fault analysis, see Paul M. Anderson, Analysis of Faulted Power Systems, New York: IEEE Press, 1995. Also see Chapters 9 and 10 of Power System Analysis by C.A. Gross, New York: Wiley, 1986.

5 Computational Methods for Electric Power Systems 5.1

Power Flow.......................................................................................... 5-1

5.2

Optimal Power Flow........................................................................ 5-13

5.3

State Estimation............................................................................... 5-25

Mariesa L. Crow Missouri University of Science and Technology

Admittance Matrix • Newton–Raphson Method

Steepest Descent Algorithm • Limitations on Independent Variables • Limitations on Dependent Variables

Electric power systems are some of the largest human-made systems ever built. As with many physical systems, engineers, scientists, economists, and many others strive to understand and predict their complex behavior through mathematical models. The sheer size of the bulk power transmission system forced early power engineers to be among the first to develop computational approaches to solving the equations that describe them. Today’s power system planners and operators rely very heavily on the computational tools to assist them in maintaining a reliable and secure operating environment. The various computational algorithms were developed around the requirements of power system operation. The primary algorithms in use today are the power flow (also known as load flow), optimal power flow (OPF), and state estimation. These are all “steady-state” algorithms and are built up from the same basic approach to solving the nonlinear power balance equations. These particular algorithms do not explicitly consider the effects of time-varying dynamics on the system. The fields of transient and dynamic stability require a large number of algorithms that are significantly different from the powerflow-based algorithms and will therefore not be covered in this chapter.

5.1 Power Flow The underlying principle of a power flow problem is that given the system loads, generation, and network configuration, the system bus voltages and line flows can be found by solving the nonlinear power flow equations. This is typically accomplished by applying Kirchoff’s law at each power system bus throughout the system. In this context, Kirchoff’s law can be interpreted as the sum of the powers entering a bus must be zero, or that the power at each bus must be conserved. Since power is comprised of two components, active power and reactive power, each bus gives rise to two equations—one for active power and one for reactive power. These equations are known as the power flow equations:

∑V Y cos(θ − θ − φ )

0 = ∆Pi = Piinj − Vi

0 = ∆Qi = Qiinj − Vi

j ij

i

j

(5.1)

ij

j =1

N

∑V Y sin(θ − θ − φ ) j ij

j =1

i

j

ij

i = 1,…, N

(5.2) 5-1

5-2

Power Systems

where Piinj and Qiinj are the active power and reactive power injected at the bus i, respectively. Loads are modeled by negative power injection. The values Vi and Vj are the voltage magnitudes at bus i and bus j, respectively. The values θi and θj are the corresponding phase angles. The value Yij ∠φij is the (ij)th element of the network admittance matrix Ybus. The constant N is the number of buses in the system. The updates ΔPi and ΔQi of Equations 5.1 and 5.2 are called the mismatch equations because they give a measure of the power difference, or mismatch, between the calculated power values, as functions of voltage and phase angle, and the actual injected powers. The formulation in Equations 5.1 and 5.2 is called the polar formulation of the power flow equations. If Yij ∠φij is instead given by the complex sum gij + jbij, then the power flow equations may be written in rectangular form as

0 = ∆Pi = Piinj − Vi

N

∑V ( g cos(θ − θ ) + b sin(θ − θ )) j

ij

i

j

ij

i

(5.3)

j

j =1

0 = ∆Qi = Qiinj − Vi

N

∑V ( g sin(θ − θ ) − b cos(θ − θ )) j

ij

i

j

ij

i

j

j =1

i = 1,…, N

(5.4)

There are, at most, 2N equations to solve. This number is then further reduced by removing one power flow equation for each known voltage (at voltage controlled buses) and the swing bus angle. This reduction is necessary since the number of equations must equal the number of unknowns in a fully determined system. In either case, the power flow equations are a system of nonlinear equations. They are nonlinear in both the voltage and phase angle. Once the number and form of the nonlinear power flow equations have been determined, the Newton–Raphson method may be applied to numerically solve the power flow equations.

5.1.1 Admittance Matrix The first step in solving the power flow equations is to obtain the admittance matrix Y for the system. The admittance matrix of a passive network (a network containing only resistors, capacitors, and inductors) may be found by summing the currents at every node in the system. An arbitrary bus i in the transmission network is shown in Figure 5.1. The bus i can be connected to any number of other buses in the system through transmission lines. Each transmission line between buses i and j is represented by a π-circuit with series impedance Rij + jXij where Rij is the per unit resistance of the transmission line. The per unit line reactance is Xij = ωs Lij where ωs is the system base frequency and Lij is the inductance of the line. In the π-circuit, the line-charging capacitance is represented by two lumped admittances jBij/2 placed at each end of the transmission line. Note that although each transmission-line parameter is in per unit and therefore a unitless quantity, the impedance is given in Ω/Ω whereas the admittance is given as Ω/Ω. The series impedances can also be represented by their equivalent admittance value where yij ∠φij =

1 Rij + jXij

Summing the currents at node i yields

I i ,inj = I i1 + Ii 2 + … + IiN + Ii10 + Ii 20 + … + IiN 0

(5.5)

5-3

Computational Methods for Electric Power Systems

i

1

Ri1

jXi1 jBi1

jBi1 2

2

jXi2

2

Ri2

jBi 2

jBi 2

2

2 N

RiN

jXiN jBiN

jBiN

2

2

FIGURE 5.1 Bus i in a power transmission network.

−V −V −V V V V ( ) ( ) ( ) … + + + = i

1

Ri1 + jXi1

i

i

2

Ri 2 + jXi 2

RiN + jXiN

j Bi1 + j Bi 2 + … + j BiN +V 2 2 2

(

)

(

iN

)

(5.6)

(

)

i − V 1 yi1∠φi1 + V i − V 2 yi 2 ∠φi 2 + + V i − V N yiN ∠φiN = V B B B i j i1 + j i 2 + + j iN +V 2 2 2

(5.7)

i = Vi ∠θi . where V Gathering the voltages yields B B B 1yi1∠φi1 − V 2 yi 2 ∠φi 2 + + V i yi1∠φi1 + yi 2 ∠φi 2 Z + + yiN ∠φiN r + j i1 + j i 2 + + j iN I i ,inj = −V 2 2 2

N yiN ∠φiN +−V

(5.8)

By noting that there are N buses in the system, there are N equations similar to Equation 5.8. These can be represented in matrix form

I1,inj Y11 I2,inj − y21∠φ21 = Ii ,inj − y1i ∠φi1 I N ,inj − y N 1∠φ N 1

− y12 ∠φ12 Y22 − yi 2 ∠φi 2 − y N 2 ∠φ N 2

… … … …

− y1i ∠φ1i − y2i ∠φ2i Yii − y Ni ∠φ Ni

… … … …

1 − y1N ∠φ1N V − y2 N ∠φ2 N V 2 1 − yiN ∠φiN V N YNN V

(5.9)

5-4

Power Systems

where B B B Yii = yi1∠φi1 + yi 2 ∠φi 2 + + yiN ∠φiN + j i1 + j i 2 + + j iN 2 2 2

(5.10)

The matrix relating the injected currents vector to the bus voltage vector is known as the system admittance matrix and is commonly represented as Y. A simple procedure for calculating the elements of the admittance matrix is Y(i,j) negative of the admittance between buses i and j Y(i,i) sum of all admittances connected to bus i

noting that the line-charging values are shunt admittances and are included in the diagonal elements.

Example 5.1 Find the admittance matrix for the line data given by i

j

Rij

Xij

Bij

1 1 2 3 3

2 5 3 4 5

0.027 0.014 0.012 0.025 0.017

0.32 0.18 0.13 0.25 0.20

0.15 0.10 0.12 0.00 0.08

Solution The first step in calculating the admittance matrix is to calculate the off-diagonal elements first. Note that in a passive network, the admittance matrix is symmetric and Y(i,j) = Y(j,i). The off-diagonal elements are calculated as the negative of the series admittance of each line. Therefore

Y (1, 2) = Y (2,1) =

−1 = 3.1139∠94.82° 0.027 + j0.32

Y (1,5) = Y (5,1) =

−1 = 5.5388∠94.45° 0.014 + j0.18

Y (2,3) = Y (3,2) =

−1 = 7.6597∠95.27° 0.012 + j0.13

Y (3,4) = Y (4,3) =

−1 = 3.9801∠95.71° 0.025 + j0.25

Y (3,5) = Y (5,3) =

−1 = 4.9820∠94.86° 0.017 + j0.20

The diagonal elements are calculated as the sum of all admittances connected to each bus.

Computational Methods for Electric Power Systems

5-5

Therefore Y (1,1) =

1 1 0.15 0.10 + +j +j = 8.5281∠−85.35° 0.027 + j0.32 0.014 + j0.18 2 2

Y (2,2) =

1 1 0.15 0.12 + +j +j = 10.6392∠−84.79° 0.027 + j0.32 0.012 + j0.13 2 2

Y (3,3) =

1 1 1 + + 0.012 + j0.13 0.025 + j0.25 0.017 + j0.20 +j

0.08 0.12 +j = 16.5221∠− 84.71° 2 2

Y ( 4, 4) =

1 = 3.9801∠−84.29° 0.025 + j0.25

Y (5,5) =

0.08 1 +j = 10.4311∠−85.32° 0.017 + j0.20 2

5.1.2 Newton–Raphson Method The most common approach to solving the power flow equations is to use the iterative Newton–Raphson method. The Newton–Raphson method is an iterative approach to solving continuous nonlinear equations in the form

f1(x1 , x2 ,…, xn ) f 2 (x1 , x2 ,…, xn ) F (x ) = =0 fn (x1 , x2 ,…, xn )

(5.11)

An iterative approach is one in which an initial guess (x 0) to the solution is used to create a sequence x 0, x1, x2,… that (hopefully) converges arbitrarily close to the desired solution vector x* where F(x*) = 0. The Newton–Raphson method for n-dimensional systems is given as

x k +1 = x k − [J (x k )]−1 F (x k )

where x1 x2 x = x3 xn

f1(x k ) k f 2 ( x ) F (x k )= f 3 (xx k ) k fn (x )

(5.12)

5-6

Power Systems

and the Jacobian matrix [J(xk)] is given by

∂f1 ∂x 1 ∂f 2 ∂x1 k [ J (x )] = ∂f 3 ∂x1 ∂f n ∂x1

∂f1 ∂x 2 ∂f 2 ∂x 2 ∂f 3 ∂x 2 ∂f n ∂x 2

∂f1 ∂x 3 ∂f 2 ∂x 3 ∂f 3 ∂x 3 ∂f n ∂x 3

… … … …

∂f 2 ∂x n ∂f 2 ∂x n ∂f 3 ∂x n ∂f n ∂x n

Typically, the inverse of the Jacobian matrix [J(xk)] is not found directly, but rather through LU factorization. Convergence is typically evaluated by considering the norm of the function F (x k ) < ε

(5.13)

Note that the Jacobian is a function of xk and is therefore updated every iteration along with F(xk). In this formulation, the vector F(x) is the set of power flow equations and the unknown x is the vector of voltage magnitudes and angles. It is common to arrange the Newton–Raphson equations by phase angle followed by the voltage magnitudes as

J J

1 3

J 2 J 4

∆δ1 ∆P1 ∆P ∆δ2 2 ∆δ3 ∆P3 ∆PN ∆δ N = − ∆V1 ∆Q1 ∆Q ∆V2 2 ∆Q ∆V3 3 ∆Q ∆QN N

(5.14)

where ∆δ i = δ ik +1 − δ ik ∆Vi = Vik +1 − Vik

and ΔPi and ΔQi are as given in Equations 5.1 and 5.2 and are evaluated at δk and V k. The Jacobian is typically divided into four submatrices, where

J J

1 3

∂∆P J 2 ∂δ = J 4 ∂∆Q ∂δ

∂∆P ∂V ∂∆Q ∂V

(5.15)

5-7

Computational Methods for Electric Power Systems

Each submatrix represents the partial derivatives of each of the mismatch equations with respect to each of the unknowns. These partial derivatives yield eight types—two for each mismatch equation, where one is for the diagonal element and the other is for off-diagonal elements. The derivatives are summarized as ∂∆Pi = Vi ∂δ i

N

∑V Y sin(δ − δ − φ ) + V Y sin φ j ij

i

j

ij

∂∆Pi =− ∂Vi

∂∆Qi =− ∂Vi

(5.17)

j ij

i

j

ij

i ii

(5.18)

ii

i =1

(5.19)

N

∑V Y cos(δ − δ − φ ) + V Y cos φ j ij

i

j

ij

i

2

ii

j =1

∂∆Qi = VV i jYij cos(δ i − δ j − φij ) ∂δ j

(5.16)

ii

∑V Y cos(δ − δ − φ ) − VV cos φ

∂∆Qi = −Vi ∂δ i

ii

N

∂∆Pi = −VY i ij cos(δ i − δ j − φij ) ∂V j

2

j =1

∂∆Pi = −VV i jYij sin(δ i − δ j − φij ) ∂δ j

i

(5.20)

ii

(5.21)

N

∑V Y sin(δ − δ − φ ) + VY sin φ j ij

i

j

ij

i ii

j =1

(5.22)

ii

∂∆Qi = −VY i ij sin(δ i − δ j − φij ) ∂V j

(5.23)

A common modification to the power flow solution is to replace the unknown update ΔVi by the normalized value ΔVi/Vi. This formulation yields a more symmetric Jacobian matrix as the Jacobian submatrices J2 and J4 are now multiplied by Vi to compensate for the scaling of ΔVi by Vi. All partial derivatives of each submatrix then become quadratic in voltage magnitude. The Newton–Raphson method for the solution of the power flow equations is relatively straightforward to program since both the function evaluations and the partial derivatives use the same expressions. Thus it takes little extra computational effort to compute the Jacobian once the mismatch equations have been calculated. Example 5.2 Find the voltage magnitudes, phase angles, and line flows for the small power system shown in Figure 5.2 with the following system: Parameters in Per Unit Bus

Type

V

Pgen

Qgen

Pload

Qload

1 2 3

Swing PV PQ

1.02 1.00 —

— 0.5 0.0

— — 0.0

0.0 0.0 1.2

0.0 0.0 0.5

5-8

Power Systems

1

2

3 PL3 + jQ L3

FIGURE 5.2 Example power system. I

j

Rij

Xij

Bij

1 1 2

2 3 3

0.02 0.01 0.01

0.3 0.1 0.1

0.15 0.1 0.1

Solution Calculating the admittance matrix for this system yields

Ybus

13.1505∠− 84.7148° = 3.326∠95.7106° 9.9504 ∠95.7106°

9.9504∠95.7106° 9.9504 ∠95.7106° 19.8012∠84.2606°

3.3260∠93.8141° 13.1505∠− 84.7148° 9.9504 ∠95.7106°

(5.24)

By inspection, this system has three unknowns: δ2, δ3, and V3; thus, three power flow equations are required. These power flow equations are 3

0 = ∆P2 = 0.5 − V2

∑V Y cos(δ − δ − θ ) j ij

j

2

ij

(5.25)

j =1 3

0 = ∆P3 = −1.2 − V3

∑V Y cos(δ − δ − θ ) j ij

3

j

ij

(5.26)

j =1 3

0 = ∆Q3 = −0.5 − V3

∑V Y sin(δ − δ − θ ) j ij

3

j

ij

(5.27)

j =1

Substituting the known quantities for V1 = 1.02, V2 = 1.00, and δ1 = 0 and the admittance matrix quantities yields

(5.28)

(1.02)(3.3260)cos(δ 2 − 0 − 93.8141°) ∆P2 = 0.5 − (1.00) +(1.00)(13.1505)cos(δ 2 − δ 2 + 84.7148°) +(V3 )(9.9504)cos(δ 2 − δ 3 − 95.7106°)

(5.29)

(1.02)(9.9504)cos(δ 2 − 0 − 95.7106°) ∆P3 = −1.2 − (V3 ) +(1.00)(9.9504)cos(δ 3 − δ 2 − 95.7106°) +(V3 )(19.8012)cos(δ 3 − δ 3 − 84.2606°)

5-9

Computational Methods for Electric Power Systems

(1.02)(9.9504)sin(δ 3 − 0 − 95.7106°) ∆Q3 = −0.5 − (V3 ) +(1.00)(9.9504)sin(δ 3 − δ 2 − 95.7106°) +(V3 )(19.8012)sin(δ 3 − δ 3 + 84.2606°)

(5.30)

The Newton–Raphson iteration for this system is then given by

∂∆P2 ∂∆P2 ∂∆P2 ∂δ ∂δ ∂V 2 3 3 ∆δ ∆P2 2 ∂∆P3 ∂∆P3 ∂∆P3 ∂δ ∂δ ∂V ∆δ3 = − ∆P3 2 3 3 ∆V3 ∆Q3 ∂∆Q3 ∂∆Q3 ∂∆Q3 ∂δ 2 ∂δ 3 ∂V3

where ∂∆P2 = 3.3925 sin(δ 2 − 93.8141°) ∂δ 2

+ 9.9504V3 sin(δ 2 − δ 3 − 94.7106°)

∂∆P2 = −9.9504V3 sin(δ 2 − δ 3 − 95.7106°) ∂δ 3

∂∆P2 = −9.9504 cos(δ 2 − δ 3 − 95.7106°) ∂V3

∂∆P3 = −9.9504V3 sin(δ 3 − δ 2 − 95.7106°) ∂δ 2 ∂∆P3 = 10.1494V3 sin(δ 3 − 95.7106°) ∂δ 3 + 9.9504V3 sin(δ 3 − δ 2 − 95.7106°) ∂∆P3 = −10.1494 cos(δ 3 − 95.7106°) ∂V3 − 9.9504 cos(δ 3 − δ 2 − 95.7106°)

− 39.6024V3 cos(84.2606° ) ∂∆Q3 = 9.9504V3 cos(δ 3 − δ 2 − 95.7106°) ∂δ 2 ∂∆Q3 = −10.1494V3 cos(δ 3 − 95.7106°) ∂δ 2

−9.9504V3 cos(δ 3 − δ 2 − 95.7106°)

(5.31)

5-10

Power Systems ∂∆Q3 = −10.1494 sin(δ 3 − 95.7106°) ∂V3 − 9.9504 sin(δ 3 − δ 2 − 95.7106°) − 39.6024V3 sin(84.2606°)

One of the underlying assumptions of the Newton–Raphson iteration is that the higher order terms of the Taylor series expansion upon which the iteration is based are negligible only if the initial guess is sufficiently close to the actual solution to the nonlinear equations. Under most operating conditions, the voltages throughout the power system are within ±10% of the nominal voltage and therefore fall in the range 0.9 ≤ Vi ≤ 1.1 per unit. Similarly, under most operating conditions the phase angle differences between adjacent buses are typically small. Thus if the swing bus angle is taken to be zero, then all phase angles throughout the system will also be close to zero. Therefore in initializing a power flow, it is common to choose a “flat start” initial condition. That is, all voltage magnitudes are set to 1.0 per unit and all angles are set to zero. Iteration 1 Evaluating the Jacobian and the mismatch equations at the flat start initial conditions yields −13.2859 [ J ] = 9.9010 −0.9901 0

9.9010 −20.000 2.0000

0.9901 −1.9604 −19.4040

∆P20 0.5044 ∆P30 = −1.1802 ∆Q30 −0.2020

Solving

0

∆δ12 ∆P20 [ J ] ∆δ13 = ∆P30 ∆V31 ∆Q30

by LU factorization yields

∆δ12 −0.0096 ∆δ13 = −0.0621 ∆V31 −0.0163

Therefore δ12 = δ 02 + ∆δ12 = 0 − 0.0096 = −0.0096 δ13 = δ 03 + ∆δ13 = 0 − 0.0621 = −0.0621

V31 = V30 + ∆V31 = 1 − 0.0163 = 0.9837

5-11

Computational Methods for Electric Power Systems

Note that the angles are given in radians and not degrees. The error at the first iteration is the largest absolute value of the mismatch equations, which is

ε1 = 1.1802

One quick check of this process is to note that the voltage update V31 is slightly less than 1.0 per unit, which would be expected given the system configuration. Note also that the diagonals of the Jacobian are all equal or greater in magnitude than the off-diagonal elements. This is because the diagonals are summations of terms, whereas the off-diagonal elements are single terms. Iteration 2 Evaluating the Jacobian and the mismatch equations at the updated values δ12, δ13, and V31 yields −13.1597 [ J ] = 9.6747 −1.4845 1

9.7771 −19.5280 3.0929

0.4684 −0.7515 −18.9086

∆P21 0.0074 ∆P31 = −0.0232 ∆Q31 −0.0359

Solving for the update yields

∆δ 22 −0.0005 ∆δ23 = −0.0014 ∆V32 −0.0021

and

δ 22 −0.0101 δ23 = −0.0635 V32 0.9816

where

ε 2 = 0.0359

Iteration 3

Evaluating the Jacobian and the mismatch equations at the updated values δ 22, δ 23, and V32 yields −13.1392 [ J ] = 9.6530 −1.4894 2

9.7567 −19.4831 −3.1079

∆P20 0.1717 0 −4 ∆P3 = −0.5639 × 10 0 −0.9084 ∆Q3

0.4600 −0.7213 −18.8300

5-12

Power Systems

Solving for the update yields

∆δ 22 −0.1396 ∆δ23 = −0.3390 × 10 −5 ∆V32 −0.5273

and

δ 32 −0.0101 δ23 = −0.0635 V33 0.9816

where

ε3 = 0.9084 × 10−4

At this point, the iterations have converged since the mismatch is sufficiently small and the values are no longer changing significantly. The last task in power flow is to calculate the generated reactive powers, the swing bus active power output and the line flows. The generated powers can be calculated directly from the power flow equations: Piinj = Vi

N

∑V Y cos(θ − θ − φ ) j ij

i

j

ij

j =1

Qiinj = Vi

N

∑V Y sin(θ − θ − φ ) j ij

i

j

ij

j =1

Therefore

Pgen,1 = P1inj = 0.7087 Q gen,1 = Q1inj = 0.2806 Q gen,2 = Q2inj = −0.0446 The total active power losses in the system are the difference between the sum of the generation and the sum of the loads, in this case:

Ploss =

∑P − ∑P gen

load

= 0.7087 + 0.5 − 1.2 = 0.0087 pu

(5.32)

The line losses for line i–j are calculated at both the sending and receiving ends of the line. The apparent power leaving bus i to bus j on line i–j is

Bij Sij = Vi ∠δ i Iij + j Vi ∠θi 2

* (5.33)

Vi ∠θi − V j ∠θ j Bij + j Vi ∠θi = Vi ∠δ i 2 Rij + jXij

* (5.34)

5-13

Computational Methods for Electric Power Systems

and the power received at bus j from bus i on line i–j is * Bij S ji = V j ∠δ j I ji − j V j ∠θ j 2

(5.35)

Thus 2 Pij = Re{Sij } = VV i jYij cos(δ i − δ j − φij ) − Vi Yij cos φij

Bij 2 Qij = Im{Sij } = VV i jYij sin(δ i − δ j − φij ) + Vi Yij sin φij − 2

(5.36) (5.37)

Similarly, the powers Pji and Q ji can be calculated. The active power loss on any given line is the difference between the active power sent from bus i and the active power received at bus j. Calculating the reactive power losses is more complex since the reactive power generated by the line-charging (shunt capacitances) must also be included.

5.2 Optimal Power Flow The basic objective of the OPF is to find the values of the system state variables and/or parameters that minimize some cost function of the power system. The types of cost functions are system dependent and can vary widely from application to application and are not necessarily strictly measured in terms of dollars. Examples of engineering optimizations can range from minimizing • • • •

Active power losses Particulate output (emissions) System energy Fuel costs of generation

to name a few possibilities. The basic formulation of the OPF can be represented as minimizing a defined cost function subject to any physical or operational constraints of the system: minimize f ( x , u) x ∈ R n n

u ∈R

(5.38)

subject to

g (x , u) = 0 equality constraints

(5.39)

h(x , u) = 0 inequality constraints

(5.40)

where x is the vector of system states u is the vector of system parameters The basic approach is to find the vector of system parameters that when substituted into the system model will result in the state vector x that minimizes the cost function f (x,u). In an unconstrained system, the usual approach to minimizing the cost function is to set the function derivatives to zero and then solve for the system states from the set of resulting equations. In the

5-14

Power Systems

majority of applications, however, the system states at the unconstrained minimum will not satisfy the constraint equations. Thus, an alternate approach is required to find the constrained minimum. One approach is to introduce an additional set of parameters λ, frequently known as Lagrange multipliers, to impose the constraints on the cost function. The augmented cost function then becomes minimize f (x , u) − λg (x , u)

(5.41)

The augmented function in Equation 5.41 can then be minimized by solving for the set of states that result from setting the derivatives of the augmented function to zero. Note that the derivative of Equation 5.41 with respect to λ effectively enforces the equality constraint of Equation 5.39.

Example 5.3 Minimize 1 C : (x 2 + y 2 ) 2

(5.42)

subject to the following constraint: 2x − y = 5

Solution

Note that the function to be minimized is the equation for a circle. The unconstrained minimum of this function is the point at the origin with x = 0 and y = 0, which defines a circle with a radius of zero length. However, the circle must also intersect the line defined by the constraint equation; thus, the constrained circle must have a nonzero radius. The augmented cost function becomes

1 C * : (x 2 + y 2 ) − λ(2 x − y − 5) 2

(5.43)

where λ represents the Lagrange multiplier. Setting the derivatives of the augmented cost function to zero yields the following set of equations:

0=

∂C * = x − 2λ ∂x

0=

∂C * = y+λ ∂y

0=

∂C * = 2x − y − 5 ∂λ

Solving this set of equations yields [x y λ]T = [2 −1 1]T. The cost function of Equation 5.42 evaluated at the minimum of the augmented cost function is

C:

1 5 [(2)2 + (−1)2 ] = 2 2

5-15

Computational Methods for Electric Power Systems If there is more than one equality constraint (i.e., if g(x, u) of Equation 5.39 is a vector of functions) then λ becomes a vector of multipliers and the augmented cost function becomes C * : f (x , u) − [λ]T g (x , u)

(5.44)

where the derivatives of C* become

∂C * = 0 = g ( x , u) ∂λ

T ∂C * ∂f ∂g = 0 = − [λ] ∂x ∂x ∂x

T ∂C * ∂f ∂g = 0 = = [λ] ∂u ∂u ∂u

(5.45)

(5.46)

(5.47)

Note that for any feasible solution, Equation 5.45 is satisfied, but the feasible solution may not be the optimal solution that minimizes the cost function. In this case, [λ] can be obtained from Equation 5.46 and then only ∂C * ∂u ≠ 0

This vector can be used as a gradient vector [∇C], which is orthogonal to the contour of constant values of the cost function C. Thus, −1

∂g T ∂f [λ] = ∂x ∂x

(5.48)

T ∂C * ∂f ∂g = − [λ] ∇C = ∂u ∂u ∂u

(5.49)

which gives

∂f ∂g = − ∂u ∂u

T

−1

∂g T ∂f ∂x ∂x

(5.50)

This relationship provides the foundation of the optimization method known as the steepest descent algorithm.

5.2.1 Steepest Descent Algorithm

1. Let k = 0. Guess an initial vector uk = u0. 2. Solve the (possibly nonlinear) system of Equation 5.45 for a feasible solution x. 3. Calculate Ck+1 and ∇Ck+1 from Equation 5.50. If ∥Ck+1 − Ck∥ is less than some predefined tolerance, stop. 4. Calculate the new vector uk+1 = uk − γ∇C, where γ is a positive number, which is the user-defined “stepsize” of the algorithm. 5. k = k + 1. Go to step 2.

5-16

Power Systems

In the steepest descent method, the u vector update direction is determined at each step of the algorithm by choosing the direction of the greatest change of the augmented cost function C*. The direction of steepest descent is perpendicular to the tangent of the curve of constant cost. The distance between adjustments is analogous to the stepsize γ of the algorithm. Thus the critical part of the steepest descent algorithm is the choice of γ. If γ is chosen small, then convergence to minimum value is more likely, but may require many iterations, whereas a large value of γ may result in oscillations about the minimum. Example 5.4 Minimize

C : x12 + 2 x22 + u 2 = f (x1 , x2 , u)

(5.51)

subject to the following constraints:

0 = x12 − 3x2 + u − 3

(5.52)

0 = x1 + x2 − 4u − 2

(5.53)

Solution To find ∇C of Equation 5.50, the following partial derivatives are required: ∂f ∂u = 2u ∂f 2 x1 ∂x = 4 x 2 T

∂g ∂u = [1 − 4]

∂g 2 x1 ∂x = 1

−3 1

yielding ∂f ∂g ∇C = − ∂u ∂u

T

−1

∂g T ∂f ∂x ∂x

2 x 1 = 2u u − [1 − 4] 1

T −3 1

−1

2x1 4 x2

Iteration 1 Let u = 1, γ = 0.05, and choose a stopping criterion of ε = 0.0001. Solving f or x1 and x2 yields two values for each with a corresponding cost function: x1 = 1.7016 x2 = 0.2984

x1 = −4.7016 x2 = 6.7016

f = 4.0734 f = 23.2828

5-17

Computational Methods for Electric Power Systems The first set of values leads to the minimum cost function, so they are selected as the operating solution. Substituting x1 = 1.7016 and x2 = 0.2984 into the gradient function yields ∇C = 10.5705 and the new value of u becomes u(2) = u(1) − γ∆C = 1 − (0.05)(10.5705)

= 0.4715

Iteration 2 With u = 0.4715, solving for x1 and x2 again yields two values for each with a corresponding cost function: x1 = 0.6062 x2 = −0.7203

f = 1.6276

x1 = −3.6062 x2 = 3.4921 f = 14.2650

The first set of values again leads to the minimum cost function, so they are selected as the operating solution. The difference in cost functions is

C (1) − C (2) = | 4.0734 − 1.6276 | = 2.4458

which is greater than the stopping criterion. Substituting these values into the gradient function yields ∇C = 0.1077 and the new value of u becomes u(3) = u(2) − γ∆C = 0.4715 − (0.05)(0.1077)

= 0.4661

Iteration 3 With u = 0.4661, solving for x1 and x2 again yields two values for each with a corresponding cost function: x1 = 0.5921 x2 = −0.7278

x1 = 3.5921 x2 = 3.4565

f = 1.6271 f = 14.1799

The first set of values again leads to the minimum cost function, so they are selected as the operating solution. The difference in cost functions is

C (2) − C (3) = | 1.6276 − 1.6271 | = 0.005

which is greater than the stopping criterion. Substituting these values into the gradient function yields ∇C = 0.0541 and the new value of u becomes u( 4) = u(3) − γ∆C = 0.4661 − (0.05)(0.0541)

= 0.4634

5-18

Power Systems Iteration 4 With u = 0.4634, solving for x1 and x2 again yields two values for each with a corresponding cost function: x1 = 0.5850 x2 = −0.7315

x1 = 3.5850 x2 = 3.4385

f = 1.6270 f = 14.1370

The first set of values again leads to the minimum cost function, so they are selected as the operating solution. The difference in cost functions is

C (3) − C ( 4) = | 1.6271 − 1.6270 | = 0.001

which satisfies the stopping criterion. Thus, the values x1 = 0.5850, x2 = −0.7315, and u = 0.4634 yield the minimum cost function f = 1.6270.

Many power system applications, such as the power flow, offer only a snapshot of the system operation. Frequently, the system planner or operator is interested in the effect that making adjustments to the system parameters will have on the power flow through lines or system losses. Rather than making the adjustments in a random fashion, the system planner will attempt to optimize the adjustments according to some objective function. This objective function can be chosen to minimize generating costs, reservoir water levels, or system losses, among others. The OPF problem is to formulate the power flow problem to find system voltages and generated powers within the framework of the objective function. In this application, the inputs to the power flow are systematically adjusted to maximize (or minimize) a scalar function of the power flow state variables. The two most common objective functions are minimization of generating costs and minimization of active power losses. The time frame of OPF is on the order of minutes to 1 h; therefore it is assumed that the optimization occurs using only those units that are currently on-line. The problem of determining whether or not to engage a unit, at what time, and for how long is part of the unit commitment problem and is not covered here. The minimization of active transmission losses saves both generating costs and creates a higher generating reserve margin. Example 5.5 Consider again the three machine system of Example 5.2 except that bus 3 has been converted to a generator bus with a voltage magnitude of 1.0 pu. The cost functions of the generators are C1: P1 + 0.0625P12 $/h C2 : P2 + 0.0125P22 $/h

C3 : P3 + 0.0250P32 $/h

Find the optimal generation scheduling of this system. Solution Following the steepest descent procedure, the first step is to develop an expression for the gradient ∇C, where

∂f ∂g ∇C = − ∂u ∂u

T

−1

∂g T ∂f ∂x ∂x

(5.54)

5-19

Computational Methods for Electric Power Systems where f is the sum of the generator costs:

f : C1 + C2 + C3 = P1 + 0.0625P12 + P2 + 0.0125P22 + P3 + 0.0250P32

g is the set of load flow equations: 3

g 1 : 0 = P2 − PL2 − V2

∑VY

i 2i

cos(δ 2 − δ i − φ2i )

i =1 3

g 2 : 0 = P3 − PL3 − V3

∑VY

i 3i

cos(δ 3 − δ i − φ3i )

i =1

where P Li denotes the active power load at bus i, the set of inputs u is the set of independent generation settings: P P

2 3

u=

and x is the set of unknown states δ δ

2 3

x=

The generator setting P1 is not an input because it is the slack bus generation and cannot be independently set. From these designations, the various partial derivatives required for ∇C can be derived: ∂g 1 ∂u = 1

∂ g1 ∂g ∂ δ 2 = ∂x ∂ g 2 ∂ δ2

(5.55)

∂ g1 ∂ δ3 ∂g2 ∂ δ3

(5.56)

where

∂ g1 ∂ δ2

∂ g1 ∂ δ3

∂g2 ∂ δ2

∂g2

∂ δ3

= V2 (V1Y12 sin(δ 2 − δ1 − φ21 ) + V3Y13 sin(δ 2 − δ 3 − φ23 ))

= V2V3Y32 sin(δ 2 − δ 3 − φ23 )

= V3V2Y23 sin(δ 3 − δ 2 − φ32 )

(5.58)

(5.59)

= V3 (V1Y13 sin(δ 3 − δ1 − φ31 ) + V2Y23 sin(δ 3 − δ 2 − φ32 ))

(5.57)

(5.60)

5-20

Power Systems and ∂f 1 + 0.025P2 = ∂u 1+ 0.050P 3

(5.61)

Finding the partial derivative [∂f/∂x] is slightly more difficult since the cost function is not written as a direct function of x. Recall, however, that P1 is not an input, but is actually a quantity that depends on x, i.e., P1 = V1(V1Y11 cos(δ1 − δ1 − φ11 )

+ V2Y12 cos(δ1 − δ 2 − φ12 ) + V3Y13 cos(δ1 − δ 3 − φ13))

(5.62)

Thus, using the chain rule,

∂f ∂f ∂P1 ∂x = ∂P ∂x 1

(5.63)

V1V2Y12 sin(δ1 − δ 2 − φ12 ) = (1 + 0.125P1 ) V1V3Y13 sin(δ1 − δ 3 − φ13 )

(5.64)

If the initial values of P2 = 0.56 pu and P3 = 0.28 pu are used as inputs, then the power flow yields the following states: [δ2 δ3] = [0.0286−0.0185] in radians and P1 = 0.1152. Converting the generated powers to megawatt and substituting these values into the partial derivatives yields

∂g 1 ∂u = 0

0 1

∂g −13.3267 ∂x = 9.8434

∂f 15.0000 ∂u = 15.0000

−52.0136 ∂f ∂x = 15.4018 155.8040

(5.65)

9.9366 −19.9219

(5.66) (5.67)

(5.68)

which yields −0.3256 ∇C = −0.4648

(5.69)

Thus, the new values for the input generation are

P 560 −0.3256 = −γ P 280 −0.4648

2 3

With γ = 1, the updated generation is P2 = 560.3 and P3 = 280.5 MW.

(5.70)

5-21

Computational Methods for Electric Power Systems Proceeding with more iterations until the gradient is reduced to less than a user-defined value yields the final generation values for all of the generators:

112.6 P P = 560.0 MW 282.7 P

1 2 3

which yields a cost of $7664/MWh.

Often the steepest descent method may indicate that either states or inputs lie outside of their physical constraints. For example, the algorithm may result in a power generation value that exceeds the physical maximum output of the generating unit. Similarly, the resulting bus voltages may lie outside of the desired range (usually ±10% of unity). These are violations of the inequality constraints of the problem. In these cases, the steepest descent algorithm must be modified to reflect these physical limitations. There are several approaches to account for limitations and these approaches depend on whether or not the limitation is on the input (independent) or on the state (dependent).

5.2.2 Limitations on Independent Variables If the application of the steepest descent algorithm results in an updated value of input that exceeds the specified limit, then the most straightforward method of handling this violation is simply to set the input state equal to its limit and continue with the algorithm except with one less degree of freedom.

Example 5.6 Repeat Example 5.5 except that the generators must satisfy the following limitations: 80 ≤ P1 ≤ 1200 MW 450 ≤ P2 ≤ 750 MW

150 ≤ P3 ≤ 250 MW

Solution From the solution of Example 5.5, the output of generator 3 exceeds the maximum limit of 0.25 pu. Therefore after the first iteration in the previous example, P3 is set to 0.25 pu. The new partial derivatives become

∂g 0 ∂u = 1

(5.71)

∂g ∂x = same

(5.72)

∂f ∂u = [1 + 0.025P2 ]

(5.73)

∂f ∂x = same

(5.74)

5-22

Power Systems From the constrained steepest descent, the new values of generation become

117.1 P P = 588.3 MW 250.0 P

1 2 3

with a cost of $7703/MWh, which is higher than the unconstrained cost of generation of $7664/ MWh. As more constraints are added to the system, the system is moved away from the optimal operating point, increasing the cost of generation.

5.2.3 Limitations on Dependent Variables In many cases, the physical limitations of the system are imposed upon states that are dependent variables in the system description. In this case, the inequality equations are functions of x and must be added to the cost function. Examples of limitations on dependent variables include maximum line flows or bus voltage levels. In these cases, the value of the states cannot be independently set, but must be enforced indirectly. One method of enforcing an inequality constraint is to introduce a penalty function into the cost function. A penalty function is a function that is small when the state is far away from its limit, but becomes increasingly larger the closer the state is to its limit. Typical penalty functions include p(h) = e kh k > 0

p(h) = x 2ne kh n, k > 0

(5.75)

(5.76)

p(h) = ax 2ne kh + be kh n, k, a, b > 0

(5.77)

and the cost function becomes C * : C(u, x ) + λT g (u, x ) + p(h(u, x ) − h max )

(5.78)

This cost equation is then minimized in the usual fashion by setting the appropriate derivatives to zero. This method not only has the advantage of simplicity of implementation, but also has several disadvantages. The first disadvantage is that the choice of penalty function is often a heuristic choice and can vary by application. A second disadvantage is that this method cannot enforce hard limitations on states, i.e., the cost function becomes large if the maximum is exceeded, but the state is allowed to exceed its maximum. In many applications this is not a serious disadvantage. If the power flow on a transmission line slightly exceeds its maximum, it is reasonable to assume that the power system will continue to operate, at least for a finite length of time. If, however, the physical limit is the height above ground for an airplane, then even a slightly negative altitude will have dire consequences. Thus the use of penalty functions to enforce limits must be used with caution and is not applicable for all systems. Example 5.7 Repeat Example 5.5, except use penalty functions to limit the power flow across line 2–3 to 0.4 pu. Solution The power flow across line 2–3 in Example 5.5 is given by P23 = V2V3Y23 cos(δ 2 − δ 3 − φ23 ) − V22Y23 cos φ23

= 0.467 pu

(5.79)

5-23

Computational Methods for Electric Power Systems If P23 exceeds 0.4 pu, then the penalty function 2 p(h) = [1000V2V3Y23 cos(δ 2 − δ 3 − φ23 ) − 1000V22Y23 cos φ23 − 400]

(5.80)

will be appended to the cost function. The partial derivatives remain the same with the exception of [|| f/ ||x], which becomes ∂f = ∂f ∂P1 + ∂f ∂P23 ∂x ∂P ∂x ∂P ∂x 1 23

(5.81)

−V2V3Y23 sin(δ 2 − δ 3 − φ23 ) V1V2Y12 sin(δ1 − δ 2 − φ1,2 ) (5.82) = (1 + 0.125P1 ) + 2(P23 − 400) V2V3Y23 sin(δ 2 − δ 3 − φ23 ) V1V3Y13 sin(δ1 − δ 3 − φ1,3 )

Proceeding with the steepest gradient algorithm iterations yields the final constrained optimal generation scheduling:

128.5 P P = 476.2 MW 349.9 P

1 2 3

and P23 = 400 MW. The cost for this constrained scheduling is $7882/MWh, which is slightly greater than the non constrained cost.

In the case where hard limits must be imposed, an alternate approach to enforcing the inequality constraints must be employed. In this approach, the inequality constraints are added as additional equality constraints with the inequality set equal to the limit (upper or lower) that is violated. This in essence introduces an additional set of Lagrangian multipliers. This is often referred to as the dual-variable approach, because each inequality has the potential of resulting in two equalities: one for the upper limit and one for the lower limit. However, the upper and lower limit cannot be simultaneously violated; thus, out of the possible set of additional Lagrangian multipliers only one of the two will be included at any given operating point and thus the dual limits are mutually exclusive. Example 5.8 Repeat Example 5.7 using the dual-variable approach. Solution By introducing the additional equation

P23 = V2V3Y23 cos(δ 2 − δ 3 − φ23 ) − V22Y23 cos φ23 = 0.400 pu

(5.83)

to the equality constraints, an additional equation gets added to the set of g(x). Therefore an additional unknown must be added to the state vector x to yield a solvable set of equations (three equations in three unknowns). Either PG2 or PG3 can be chosen as the additional unknown. In this example, PG3 will be chosen. The new system Jacobian becomes

∂g 1 ∂ x1 ∂g ∂ g 2 ∂x = ∂ x1 ∂ g3 ∂x 1

∂ g1

∂x2

∂g2 ∂x2

∂g3 ∂x2

∂ g1 ∂x3 ∂g2 ∂x3 ∂g3 ∂ x 3

(5.84)

5-24

Power Systems where ∂ g1 ∂ x1

∂ g1

∂x2

∂ g1

∂x3 ∂g2 ∂ x1

∂g2 ∂x2

∂g2 ∂x3

∂g3 ∂ x1

∂g3 ∂x2

∂g3

∂x3

= V2 (V1Y12 sin(δ 2 − δ1 − φ21 ) + V3Y13 sin(δ 2 − δ 3 − φ23 )) = −V2V3Y32 sin(δ 2 − δ 3 − φ23 ) =0 = −V3V2Y23 sin(δ 3 − δ 2 − φ32 ) = V3 (V1Y13 sin(δ 3 − δ1 − φ31 ) + V2Y23 sin(δ 3 − δ 2 − φ32 ) =1 = −V2V3Y23 sin(δ 2 − δ 3 − φ23 ) = V2V3Y23 sin(δ 2 − δ 3 − φ23 ) =0

and

1 ∂g ∂f = ∂u 0 ; ∂u = [1 + 0.025PG 2 ] 0

Similar to Example 5.5, the chain rule is used to obtain [∂f/∂x]:

∂f ∂C ∂PG1 ∂C ∂PG 3 ∂x = ∂P ∂x + ∂P ∂x G3 G1

(5.85)

V1V2Y12 sin(δ1 − δ 2 − φ12 ) = (1 + 0.125PG1 ) V1V3Y13 sin(δ1 − δ3 − φ13 ) 0 V3V2Y32 sin(∂3 − ∂2 − φ32 ) + (1 + 0.050PG 3 ) −V3 V1Y13 sin(δ 3 − δ1 − φ31 ) + V2Y23 sin(δ 3 − δ 2 − φ32 ) 0

(

)

(5.86)

Substituting these partial derivatives into the expression for ∇C of Equation 5.54 yields the same generation scheduling as Example 5.7.

5-25

Computational Methods for Electric Power Systems

5.3 State Estimation In many physical systems, the system operating condition cannot be determined directly by an analytical solution of known equations using a given set of known, dependable quantities. More frequently, the system operating condition is determined by the measurement of system states at different points throughout the system. In many systems, more measurements are made than are necessary to uniquely determine the operating point. This redundancy is often purposely designed into the system to counteract the effect of inaccurate or missing data due to instrument failure. Conversely, not all of the states may be available for measurement. High temperatures, moving parts, or inhospitable conditions may make it difficult, dangerous, or expensive to measure certain system states. In this case, the missing states must be estimated from the rest of the measured information of the system. This process is often known as state estimation and is the process of estimating unknown states from measured quantities. State estimation gives the “best estimate” of the state of the system in spite of uncertain, redundant, and/or conflicting measurements. A good state estimation will smooth out small random errors in measurements, detect and identify large measurement errors, and compensate for missing data. This process strives to minimize the error between the (unknown) true operating state of the system and the measured states. The set of measured quantities can be denoted by the vector z, which may include measurements of system states (such as voltage and current) or quantities that are functions of system states (such as power flows). Thus,

z true = Ax

(5.87)

where x is the set of system states A is usually not square The error vector is the difference between the measured quantities z and the true quantities:

e = z − z true = z − Ax

(5.88)

Typically, the minimum of the square of the error is desired to negate any effects of sign differences between the measured and true values. Thus, a state estimator endeavors to find the minimum of the squared error, or a least squares minimization: 2

minimize e = eT ⋅ e =

m i i =1

m

∑ z − ∑a x ij

j =1

j

2

(5.89)

The squared error function can be denoted by U(x) and is given by

U (x ) = eT ⋅ e = (z − Ax )T (z − Ax )

(5.90)

= (z T − x T AT )(z − Ax ) p

(5.91)

= (z T z − z T Ax − x T AT z + x T AT Ax )

(5.92)

Note that the product zT Ax is a scalar and so it can be equivalently written as

z T Ax = (z T Ax )T = x T AT zp

Therefore the squared error function is given by

U (x ) = z T z − 2x T AT z + x T AT Ax

(5.93)

5-26

Power Systems

The minimum of the squared error function can be found by an unconstrained optimization where the derivative of the function with respect to the states x is set to zero:

∂U ( x ) = 0 − 2 AT z + 2 AT z + 2 AT Ax ∂x

(5.94)

AT Ax = AT z

(5.95)

Aˆ x = b

(5.96)

Thus, Thus, if b = ATz and Aˆ = AT A, then

This state vector x is the best estimate (in the squared error) to the system operating condition from which the measurements z were taken. The measurement error is given by

e = z meas − Ax

(5.97)

In power system state estimation, the estimated variables are the voltage magnitudes and the voltage phase angles at the system buses. The input to the state estimator is the active and reactive powers of the system, measured either at the injection sites or on the transmission lines. The state estimator is designed to give the best estimates of the voltages and phase angles minimizing the effects of the measurement errors. All instruments add some degree of error to the measured values, but the problem is how to quantify this error and account for it during the estimation process. If all measurements are treated equally in the least squares solution, then the less accurate measurements will affect the estimation as significantly as the more accurate measurements. As a result, the final estimation may contain large errors due to the influence of inaccurate measurements. By introducing a weighting matrix to emphasize the more accurate measurements more heavily than the less accurate measurements, the estimation procedure can then force the results to coincide more closely with the measurements of greater accuracy. This leads to the weighted least squares estimation: 2

m

T

minimize e = e ⋅ e =

∑w i =1

i i

z −

m

∑a x ij

j =1

j

2

(5.98)

where wi is a weighting factor reflecting the level of confidence in the measurement zi. In general, it can be assumed that the introduced errors have normal (Gaussian) distribution with zero mean and that each measurement is independent of all other measurements. This means that each measurement error is as likely to be greater than the true value as it is to be less than the true value. A zero mean Gaussian distribution has several attributes. The standard deviation of a zero mean Gaussian distribution is denoted by σ. This means that 68% of all measurements will fall within ±σ of the expected value, which is zero in a zero mean distribution. Further, 95% of all measurements will fall within ±2σ and 99% of all measurements will fall within ±3σ. The variance of the measurement distribution is given by σ2. This implies that if the variance of the measurements is relatively small, then the majority of measurements are close to the mean. One interpretation of this is that accurate measurements lead to small variance in the distribution. This relationship between accuracy and variance leads to a straightforward approach from which a weighting matrix for the estimation can be developed. With measurements taken from a particular

5-27

Computational Methods for Electric Power Systems

meter, the smaller the variance of the measurements (i.e., the more consistent they are), the greater the level of confidence in that set of measurements. A set of measurements that have a high level of confidence should have a higher weighting than a set of measurements that have a larger variance (and therefore less confidence). Therefore, a plausible weighting matrix that reflects the level of confidence in each measurement set is the inverse of the covariance matrix. Thus, measurements that come from instruments with good consistency (small variance) will carry greater weight than measurements that come from less accurate instruments (high variance). Thus, one possible weighting matrix is given by

1 σ2 1 0 W = R −1 = 0

0

0 0 1 σm2

…

1 σ22

…

0

…

(5.99)

where R is the covariance matrix for the measurements. As in the linear least squares estimation, the nonlinear least squares estimation attempts to minimize the square of the errors between a known set of measurements and a set of weighted nonlinear functions:

m

2

minimize f = e = eT ⋅ e =

∑ σ [z − h (x)] 1

2

i

i

i =1

2

(5.100)

where x ∊ Rn is the vector of unknowns to be estimated, z ∊ Rm is the vector of measurements, σi2 is the variance of the ith measurement, and h(x) is the nonlinear function vector relating x to z, where the measurement vector z can be a set of geographically distributed measurements, such as voltages and power flows In state estimation, the unknowns in the nonlinear equations are the state variables of the system. The state values that minimize the error are found by setting the derivatives of the error function to zero:

F (x ) = H Tx R −1[z − h(x )] = 0

(5.101)

where

∂h1 ∂x 1 ∂h2 H x = ∂x1 ∂hm ∂x1

∂h1 ∂x 2 ∂h2 ∂h2 ∂hm ∂x 2

… … …

∂h1 ∂x n ∂h2 ∂x n ∂hm ∂x n

(5.102)

Note that Equation 5.101 is a set of nonlinear equations that must be solved using Newton–Raphson or another iterative numerical solver. In this case, the Jacobian of F(x) is

J F (x ) = H Tx (x )R −1

∂ [z − h(x )] ∂x

= − H Tx (x )R −1H x (x )

(5.103) (5.104)

5-28

Power Systems

and the Newton–Raphson iteration becomes

[H Tx (x k )R −1H x (x k )][x k −1 − x k ] = H Tx (x k )R −1[z − h(x k )]

(5.105)

At convergence, xk + 1 is equal to the set of states that minimize the error function f of Equation 5.100. Example 5.9 The SCADA system for the power network shown in Figure 5.3 reports the following measurements and variances: zi

State

Measurement

Variance (s2)

1 2 3 4 5

V3 P13 Q21 P3 Q2

0.975 0.668 −0.082 −1.181 −0.086

0.010 0.050 0.075 0.050 0.075

Estimate the power system states. Solution The first step in the estimation process is to identify and enumerate the unknown states. In this example, the unknowns are [x1 x2 x3]T = [δ2 δ3 V3]T. After the states are identified, the next step in the estimation process is to identify the appropriate functions h(x) that correspond to each of the measurements. The nonlinear function that is being driven to zero to minimize the weighted error is

F (x ) = H Tx R −1[z − h(x )] = 0

(5.106)

where the set of z − h(x) is given by z1 − h1(x ) = V3 − x3

(

z2 − h2(x ) = P13 − V1x3Y12 cos( − x2 − φ13 ) − V12Y13 cos φ13

(

z 3 − h3 (x ) = Q21 − V2V1Y21 sin(x1 − φ21 ) + V22Y21 sin φ21

)

)

(

z 4 − h4 (x ) = P3 − x3V1Y31 cos(x2 − φ31 ) + x3V2Y32 cos(x2 − x1 − φ32 ) + x32Y33 cos φ33

)

z 5 − h5 (x ) = Q2 − V2V1Y21 sin(x1 − φ21 ) − V22Y22 sin φ22 + V2 x3Y23 sin(x1 − x2 − φ23 )

1

2

3 PL3 + jQ L3

FIGURE 5.3 Example power system.

5-29

Computational Methods for Electric Power Systems and the matrix of partial derivatives for the set of functions in Equation 5.106 is

0 0 sin( φ ) 0 V x Y − x − 1 3 13 2 13 V1V2Y21 cos( x1 − φ21 ) 0 Hx = − x3V1Y31 sin(x2 − φ31 ) − x3V2Y32 sin(x2 − x1 − φ32 ) x3V2Y21 sin(x2 − xφ32 ) V V Y cos( x − φ ) + V x Y cos(x − x − φ ) −V2 x3Y23 cos( x1 − x2 − φ23 ) 1 21 2 3 23 1 2 23 1 2 21

1 V1Y13 cos( − x2 − φ13 ) 0 × V Y cos( x − φ ) + V Y cos( x − x − φ ) + x Y cos φ 2 3 33 33 2 31 2 32 2 1 32 1 31 V2Y23 sin(x1 − x2 − φ23 )

(5.107)

The covariance matrix of the measurements is

1 0.0102 R=

1 0.0502

1 0.0752

1 0.0502

1 0.0752

(5.108)

The Newton–Raphson iteration to solve for the set of states x that minimize the weighted errors is

[H Tx (x k )R −1H x (x k )][x k −1 − x k ] = H Tx (x k )R −1[z − h(x k )]

(5.109)

Iteration 1 The initial condition for the state estimation solution is the same flat start as for the power flow equations; namely, all angles are set to zero and all unknown voltage magnitudes are set to unity. The measurement functions h(x) evaluated at the initial conditions are

1.0000 0.0202 h(x 0 ) = −0.0664 −0.0198 −0.1914

5-30

Power Systems The matrix of partials evaluated at the initial condition yields

0 0 0 H x = −0.2257 −9.9010 −1.2158

0 −10.0990 0 20.0000 0.9901

1.0000 −1.0099 0 1.9604 −9.9010

The nonlinear functions in Equation 5.106 are 0.5655 F (x 0 ) = −1.4805 −0.2250

The incremental updates for the states are −0.0119 ∆x = −0.0625 −0.0154 1

leading to the updated states

δ12 −0.0119 δ13 = −0.0625 V31 −0.9846

where δ2 and δ3 are in radians. The error at iteration 0 is

ε 0 = 1.4805

Iteration 2 The updated values are used to recalculate the Newton–Raphson iterations:

0.9846 0.6585 h(x1 ) = −0.0634 −1.1599 −0.724

The matrix of partials is

0 0 1 H x = −0.2660 −9.6864 −0.7468

0 −9.9858 0 19.5480 0.4809

1.0000 −0.3774 0 0.7715 −9.9384

Computational Methods for Electric Power Systems

5-31

The nonlinear function evaluated at the updated values yields

0.0113 F (x1 ) = −0.0258 0.0091

The incremental updates for the states are 0.0007 ∆x = −0.0008 0.0013 2

leading to the updated states

δ 22 −0.0113 2 δ 3 = −0.0633 2 0.9858 V3

The error at Iteration 1 is

ε1 = 0.0258

The iterations are obviously converging. At convergence, the states that minimize the weighted measurement errors are

−0.0113 x = −0.0633 0.9858

This concludes the discussion of the most commonly used computational methods for power system analysis. This chapter describes only the basic approaches to power flow, OPF, and state estimation. These methods have been utilized for several decades, yet improvements in accuracy and speed are constantly being proposed in the technical literature.

Power System Transients

II

Pritindra Chowdhuri

6 Characteristics of Lightning Strokes Francisco De la Rosa............................................. 6-1

7 Overvoltages Caused by Direct Lightning Strokes Pritindra Chowdhuri......................7-1

8 Overvoltages Caused by Indirect Lightning Strokes Pritindra Chowdhuri................. 8-1

9 Switching Surges Stephen R. Lambert................................................................................. 9-1

10 Very Fast Transients Juan A. Martinez-Velasco............................................................... 10-1

11 Transmission System Transients: Grounding William A. Chisholm............................11-1

Introduction • Lightning Generation Mechanism • Parameters of Importance for Electric Power Engineering • Incidence of Lightning to Power Lines • Conclusions • References

Direct Strokes to Unshielded Lines • Direct Strokes to Shielded Lines • Significant Parameters • Outage Rates by Direct Strokes • Effects of Induction for Direct Strokes • Ground Impedance and Corona under Lightning • References

Inducing Voltage • Induced Voltage • Green’s Function • Induced Voltage of a Doubly Infinite Single-Conductor Line • Induced Voltages on Multiconductor Lines • Effects of Shield Wires on Induced Voltages • Stochastic Characteristics of Lightning Strokes • Estimation of Outage Rates Caused by Nearby Lightning Strokes • Estimation of Total Outage Rates • Appendix A: Voltage Induced by Linearly Rising and Falling Return-Stroke Current • References Transmission Line Switching Operations • Series Capacitor Bank Applications • Shunt Capacitor Bank Applications • Shunt Reactor Applications • References Origin of VFT in GIS • Propagation of VFT in GIS • Modeling Guidelines and Simulation • Effects of VFT on Equipment • References

General Concepts • Material Properties • Electrode Dimensions • Self-Capacitance of Electrodes • Initial Transient Response from Capacitance • Ground Electrode Impedance: Wire over Perfect Ground • Ground Electrode Impedance: Wire over Imperfect Ground • Analytical Treatment of Complex Electrode Shapes • Numerical Treatment of Complex Electrode Shapes • Treatment of Multilayer Soil Effects • Layer of Finite Thickness over Insulator • Treatment of Soil Ionization • Design Process • Design Recommendations • Appendix A: Relevant IEEE and IEC Standards in Lightning and Grounding • References

II-1

II-2

Power System Transients

12 Transient Recovery Voltage Thomas E. McDermott....................................................... 12-1

13 Surge Arresters Thomas E. McDermott............................................................................ 13-1

14 Insulation Coordination Stephen R. Lambert................................................................. 14-1

Fault Interruption Process • Analysis Principles • TRV for Transformer-Fed Faults • TRV for Capacitor Bank Switch Opening • TRV for Line-Fed Faults • TRV forCurrent-Limiting Reactor Faults • Switchgear Tests and Standards • TRV Mitigation • References Arrester Types and Auxiliary Equipment • Ratings and Tests • Selection by TOV • Selection by Energy Rating • Arrester Modeling • Applications • References Insulation Coordination • Insulation Characteristics • Probability of Flashover • Flashover Characteristics of Air Insulation • References

Pritindra Chowdhuri received his BSc in physics and MSc in applied physics from University of Calcutta, Kolkata, India, MS in electrical engineering from Illinois Institute of Technology, Chicago, Illinois, and DEng in engineering science from Rensselaer Polytechnic Institute, Troy, New York. He has worked for General Electric Company in Pittsfield, Massachusetts, Schenectady, New York, and Erie, Pennsylvania, as a development engineer and for Los Alamos National Laboratory in Los Alamos, New Mexico, as a staff member. He joined Tennessee Technological University (TTU) in 1986 as a professor of electrical engineering in the Center for Electric Power (later Center for Energy Systems Research) and the Department of Electrical Engineering. He retired from TTU in 2005 as emeritus professor.

6 Characteristics of Lightning Strokes

Francisco De la Rosa Electric Power Systems

6.1 6.2

Introduction....................................................................................... 6-1 Lightning Generation Mechanism.................................................. 6-2

6.3

Parameters of Importance for Electric Power Engineering........6-4

First Strokes • Subsequent Strokes

Ground Flash Density • Current Peak Value • Correlation between Current and Other Parameters of Lightning

6.4 Incidence of Lightning to Power Lines........................................... 6-7 6.5 Conclusions........................................................................................6-8 References.......................................................................................................6-8

6.1 Introduction Lightning, one of the most spectacular events of Mother Nature, started to appear significantly demystified after Franklin showed its electric nature with his famous electrical kite experiment in 1752. Although a great deal of research on lightning followed Franklin’s observation, lightning continues to be a topic of considerable interest for investigation (Uman, 1969, 1987). This is particularly true for the improved design of electric power systems, since lightning-caused interruptions and equipment damage during thunderstorms stand as the leading causes of failures in the electric utility industry. It is prudent to state that in spite of the impressive amount of lightning research conducted mostly during the last 50 years, the physics of the phenomenon is not yet fully understood. The development of powerful digital recorders with enough bandwidth to capture the microstructure of lightning waveforms and the advent of digital computers and fiber optic communication along with sophisticated direction finding sensors during this period facilitated the extraordinary evolution of lightning monitoring and detection techniques. This allowed for a more accurate description of lightning and the characterization of lightning parameters that are relevant for power system protection. On a worldwide scale, most lightning currents (over 90%) are of negative polarity. However, it has to be acknowledged that in some parts of the world, mostly over the Northern Hemisphere, the fraction of positive lightning currents can be substantial. A formal assessment of the effects of positive lightning on electric power systems will require using the corresponding parameters, since they may show large deviations from negative flashes in peak current, charge, front duration, and flash duration as it will be described.

6-1

6-2

Power Systems

6.2 Lightning Generation Mechanism 6.2.1 First Strokes The wind updrafts and downdrafts that take place in the atmosphere create a charging mechanism that separates electric charges, leaving negative charge at the bottom and positive charge at the top of the cloud. As charge at the bottom of the cloud keeps growing, the potential difference between cloud and ground, which is positively charged, grows as well. This process will continue until air breakdown occurs. See Figure 6.1. The way in which a cloud-to-ground flash develops involves a stepped leader that starts traveling downward following a preliminary breakdown at the bottom of the cloud. This involves a positive pocket of charge, as illustrated in Figure 6.1. The stepped leader travels downward in steps several tens of meters in length and pulse currents of at least 1 kA in amplitude (Uman, 1969). When this leader is near ground, the potential to ground can reach values as large as 100 MV before the attachment process with one of the upward streamers is completed. Figure 6.2 illustrates the final step when the upward streamer developing from a transmission line tower intercepts the downward leader. The connection point-to-ground is not decided until the downward leader is some tens of meters above the ground plane. The downward leader will be attached to one of the growing upward streamers developing from elevated objects such as trees, chimneys, power lines, telecommunication towers, etc. It is actually under this principle that lightning protection rods work, that is, they have to be strategically located so that they can trigger upward streamers that can develop into attachment points to downward leaders approaching the protected area. For this to happen, upward streamers developing from Positive charge center on the top of the cloud + + + + ++ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + ++ + + + + + ++ +

Negative charge center on the bottom of the cloud – – – – – – – – – – – – – – – – – – – – – – Preliminary breakdown + ++ + +– – – involving a pocket – – – – – of positive charge Upward streamer from a tall mast + + + + + + + + + + + +

+ ++ +

+ + + +

FIGURE 6.1 Separation of electric charge within a thundercloud. – – – – –– – – – – – – – – – –

– – – – –– – – – – – – – – – – + +

+ + (a)

+++

+++

(b)

+++

– – – – –– – – – – – – – – – – + + +++

Downward stepped leader progression

(c)

++ +

– – – – –– – – + – – – + + – + + + +++

(d)

+++

– – – – + + + + + + + + + + + + + + + + +++

(e)

+++

Attachment process and return stroke development

FIGURE 6.2 Attachment between downward and upward leaders in a cloud-to-ground flash.

+++

6-3

Characteristics of Lightning Strokes TABLE 6.1 Lightning Current Parameters for Negative Flashes Parameters

Units

Peak current (minimum 2 kA) First strokes Subsequent strokes Charge (total charge) First strokes Subsequent strokes Complete flash Impulse charge (excluding continuing current) First strokes Subsequent strokes Front duration (2 kA to peak) First strokes Subsequent strokes Maximum di/dt First strokes Subsequent strokes Stroke duration (2 kA to half peak value on the tail) First strokes Subsequent strokes

kA

Action integral (∫ i 2 dt ) First strokes Subsequent strokes Time interval between strokes Flash duration All flashes Excluding single-stroke flashes

Sample Size

Value Exceeding in 50% of the Cases

101 135

30 12

93 122 94

5.2 1.4 7.5

90 117

4.5 0.95

89 118

5.5 1.1

92 122

12 40

90 115

75 32

91 88 133

5.5 × 104 6.0 × 103 33

C

C

μs

kA/μs

μs

A2 s ms ms

94 39

13 180

Source: Adapted from Berger, K. et al., Electra, 41, 23, 1975.

protected objects within the shielded area have to compete unfavorably with those developing from the tip of the lightning rods, which are positioned at a higher elevation. Just after the attachment process takes place, the charge that is lowered from the cloud base through the leader channel is conducted to ground as a breakdown current pulse, known as the return stroke, travels upward along the channel. The return stroke velocity is around one-third the speed of light. The median peak current value associated to the return stroke is reported to be on the order of 30 kA, with rise time and time to half values around 5 and 75 μs, respectively. See Table 6.1 adapted from Berger et al. (1975). Associated to this charge-transfer mechanism (an estimated 5 C total charge is lowered to ground through the stepped leader) are the electric and magnetic field changes that can last several milliseconds. These fields can be registered at remote distances from the channel and it is under this principle that lightning sensors work to produce the information necessary to monitor cloud-to-ground lightning.

6.2.2 Subsequent Strokes After the negative charge from the cloud base has been transferred to ground, additional charge can be made available on the top of the channel when discharges known as J and K processes take place within the cloud (Uman, 1969). This can lead to a number of subsequent strokes of lightning following the first stroke. A so-called dart leader develops from the top of the channel lowering charges, typically of 1 C, until recently believed to follow the same channel of the first stroke. Studies conducted in the past few years,

6-4

Power Systems

however, suggest that around half of all lightning discharges to earth, both single- and multiple-stroke flashes, may strike ground at more than one point, with separation between channel terminations on ground varying from 0.3 to 7.3 km and a geometric mean of 1.3 km (Thottappillil et al., 1992). Generally, dart leaders develop no branching and travel downward at velocities of around 3 × 106 m/s. Subsequent return strokes have peak currents usually smaller than first strokes but faster zero-to-peak rise times. The mean interstroke interval is about 60 ms, although intervals as large as a few tenths of a second can be involved when a so-called continuing current flows between strokes (this happens in 25%–50% of all cloud-to-ground flashes). This current, which is on the order of 100 A, is associated to charges of around 10 C and constitutes a direct transfer of charge from cloud to ground (Uman, 1969). The percentage of single-stroke flashes presently suggested by CIGRE of 45% (Anderson and Eriksson, 1980) are considerably higher than the following figures recently obtained from experimental results: 17% in Florida (Rakov et al., 1994), 14% in New Mexico (Rakov et al., 1994), 21% in Sri Lanka (Cooray and Jayaratne, 1994), and 18% in Sweden (Cooray and Perez, 1994).

6.3 Parameters of Importance for Electric Power Engineering 6.3.1 Ground Flash Density Ground flash density, frequently referred to as GFD or Ng, is defined as the number of lightning flashes striking ground per unit area and per year. Usually it is a long-term average value and ideally it should take into account the yearly variations that take place within a solar cycle—believed to be the period within which all climatic variations that produce different GFD levels occur. A 10-year average GFD map of the continental United States (IEEE Guide, 2005) obtained by and reproduced here with permission from Vaisala, Inc. of Tucson, Arizona, is presented in Figure 6.3. Note the considerable large GFD levels affecting remarkably the state of Florida, as well as all the southern states along the Gulf of Mexico (Alabama, Mississippi, Louisiana, and Texas). High GFD levels are also

>1

>2

>2

0.5

3

4 3

1

1

>2

2

>1

10

8 12 10>10 8 10 14 14 >14

FIGURE 6.3 Ten-year average GFD map of the United States. (Reproduced from Vaisala, Inc. of Tucson, AZ., Standards Information Network, How to protect your house and its contents from lightning, IEEE Guide for Surge Protection of Equipment Connected to AC Power and Communication Circuits, IEEE Press, New York, 2005. With permission.)

Characteristics of Lightning Strokes

6-5

observed in the southeastern states of Georgia and South Carolina. To the west side, Arizona is the only state with GFD levels as high as 8 flashes/km2/year. The lowest GFD levels ( Io and Q > Qo), thereby refining the probability estimates for the conditions under which a lightning-related failure may occur. Table 6.3 comprises relevant findings for positive flashes, negative first strokes, and negative subsequent strokes. The table entries for parameters in negative discharges provide more accurate results, due to broader bandwidth (higher frequency) recording instruments. Table entries that are not filled in (—) were not analyzed by Dellera. To illustrate the interpretation of the table, we observe that a moderate-to-high correlation is found between lightning current and all but peak rate of rise in positive flashes. Therefore, extreme heating should be expected in arcing or transient currents conducted through protective devices following insulation flashover produced by positive lightning flashes. Note that lightning parameters associated with heat are charge and action integral and that rate-of-change of lightning current is connected with inductive effects, which do not show strong from the correlation table. Similarly, overvoltages from negative lightning strokes should be associated with peak current producing large inductive overvoltages especially due to subsequent strokes because of their larger di/dt (by a factor of 3 or more) relative to first strokes. Heating effects, however, are loosely correlated with peak current, since correlation coefficient for the total charge (Qflash) is poor. It is possible that consideration

6-7

Characteristics of Lightning Strokes TABLE 6.3 Correlation between Lightning Parameters Lightning Parameter

Correlation (Correlation Coefficient)

Data Source

Positive flashes Front time (Tfront) Peak rate of rise (di/dt) Impulse charge (Qimp) Flash charge (Qflash) Impulse action integral (Wimp) Flash action integral (Wflash)

Low (0.18) Moderate (0.55) High (0.77) Moderate (0.59) — High (0.76)

Berger and Garbagnati (1984) Berger and Garbagnati (1984) Berger and Garbagnati (1984) Berger and Garbagnati (1984)

Negative first strokes Front time (Tfront) Peak rate of rise (di/dt) Impulse charge (Qimp) Flash charge (Qflash) Impulse action integral (Wimp) Flash action integral (Wflash)

Low (—) Moderate/high (—) High (0.75) Low (0.29) High (0.86) —

Weidman and Krider (1984)a Weidman and Krider (1984)a Berger and Garbagnati (1984) Berger and Garbagnati (1984) Berger and Garbagnati (1984)

Negative subsequent strokes Front time (Tfront) Peak rate of rise (di/dt)

Low (0.13) High (0.7–0.8)

Fisher et al. (1993) Fisher et al. (1993) and Leteinturier et al. (1990)b

Impulse charge (Qimp) Impulse action integral (Wimp)

Berger and Garbagnati (1984)

— —

Source: Adapted from Dellera, L., Lightning parameters for protection: And updated approach, CIGRE 97 SC33, WG33.01, 17 IWD, August 1997. a Inferred from measurements of electric fields propagated over salt water. b 30%–90% slope, which corresponds to an “average” di/dt (triggered lightning studies).

of all stroke peak currents and interstroke intervals in negative flashes, which are available in some modern LDS, will prove a more deterministic means to infer heating due to negative flashes.

6.4 Incidence of Lightning to Power Lines One of the most accepted expressions to determine the number of direct strikes to an overhead line in an open ground with no nearby trees or buildings is described by Eriksson (1987):

28h0.6 + b N = Ng 10

(6.3)

where h is the pole or tower height (m)—negligible for distribution lines b is the structure width (m) Ng is the GFD (flashes/km2/year) N is the number of flashes striking the line/100 km/year For unshielded distribution lines, this is comparable to the fault index due to direct lightning hits. For transmission lines, this is an indicator of the exposure of the line to direct strikes (the response of the line being a function of overhead ground wire shielding angle on one hand and on conductor-tower surge impedance and footing resistance on the other hand).

6-8

Power Systems

Note the dependence of the incidence of strikes to the line with height of the structure. This is important since transmission lines are several times taller than distribution lines, depending on their operating voltage level. Also important is that in the real world, power lines are to different extents shielded by nearby trees or other objects along their corridors. This will decrease the number of direct strikes estimated by Equation 6.3 to a degree determined by the distance and height of the objects. In IEEE Std. 1410-1997, a shielding factor is proposed to estimate the shielding effect of nearby objects to the line. An important aspect of this reference work is that objects within 40 m from the line, particularly if they are equal or higher than around 20 m, can attract most of the lightning strikes that would otherwise hit the line. Likewise, the same objects would produce insignificant shielding effects if located beyond 100 m from the line. On the other hand, sectors of lines extending over hills or mountain ridges may increase the number of strikes to the line. The aforementioned effects may in some cases cancel one another so that the estimation obtained form Equation 6.3 can still be valid. However, it is recommended that any assessment of the incidence of lightning strikes to a power line be performed by taking into account natural shielding and orographic conditions (terrain undulations) along the line route. This also applies when identifying troubled sectors of the line for the installation of metal oxide surge arresters to improve its lightning performance. For example, those segments of a distribution feeder crossing over hills with little natural shielding would greatly benefit from surge arrester protection. Finally, the inducing effects of lightning, also described in Anderson et al. (1984) and De la Rosa et al. (1998), have to be considered to properly estimate distribution line lightning performance or when estimating the outage rate improvement after application of any mitigation action. Lightning strokes terminating on ground close to distribution lines have the potential to develop overvoltages large enough to cause insulation flashover. Under certain conditions, like in circuits without grounded neutral, with low critical flashover voltages, high GFD levels, or located on high-resistivity terrain, the number of outages produced by close lightning can considerably surpass those due to direct strikes to the line.

6.5 Conclusions Parameters that are important for the assessment of lightning performance of power transmission and distribution lines or for evaluation of different protection methods are lightning current and GFD. The former provides a means to appraise the impact of direct hits on power lines or substations and the latter provides an indication of how often this phenomenon may occur. There are, however, other lightning parameters that can be related to the probability of insulation flashover and heating effects in surge protective devices, which are difficult to obtain from conventional lightning detection equipment. Some correlation coefficients observed between lightning peak current and other relevant parameters obtained from significant experiments in instrumented towers are portrayed in this review. Aspects like different methods available to calculate shielding failures and back flashovers in transmission lines or the efficacy of remedial measures to improve lightning performance of electrical networks are not covered in this summarizing review. Among these, overhead ground wires, metal oxide surge arresters, increased insulation, or use of wood as an arc-quenching device can only be mentioned. The reader is encouraged to further look at the suggested references or to get experienced advice for a more comprehensive understanding on the subject.

References Anderson, R.B. and Eriksson, A.J., Lightning parameters for engineering applications, Electra, 69, 65–102, March 1980. Anderson, R.B., Eriksson, A.J., Kroninger, H., Meal, D.V., and Smith, M.A., Lightning and thunderstorm parameters, IEE Lightning and Power Systems Conference Publication No. 236, London, U.K., pp. 57–61, 1984.

Characteristics of Lightning Strokes

6-9

Berger, K., Anderson, R.B., and Kroninger, H., Parameters of lightning flashes, Electra, 41, 23–37, July 1975. Berger, K. and Garbagnati, E., Lightning current parameters, results obtained in Switzerland and in Italy, in Proceedings of URSI Conference, Florence, Italy, p. 13, 1984. Cooray, V. and Jayaratne, K.P.S., Characteristics of lightning flashes observed in Sri Lanka in the tropics, J. Geophys. Res., 99, 21,051–21,056, 1994. Cooray, V. and Perez, H., Some features of lightning flashes observed in Sweden, J. Geophys. Res., 99, 10,683–10,688, 1994. De la Rosa, F., Cummins, K., Dellera, L., Diendorfer, G., Galvan, A., Huse, J., Larsen, V., Nucci, C.A., Rachidi, F., Rakov, V., Torres, H. et al., Characterization of lightning for applications in electric power systems, CIGRE Report#172, TF33.01.02, December 2000. De la Rosa, F., Nucci, C.A., and Rakov, V.A., Lightning and its impact on power systems, in Proceedings of International Conference on Insulation Coordination for Electricity Development in Central European Countries, Zagreb, Croatia, p. 44, 1998. Dellera, L., Lightning parameters for protection: An updated approach, CIGRE 97 SC33, WG33.01, 17 IWD, August 1997. Eriksson, A.J., The incidence of lightning strikes to power lines, IEEE Trans. Power Delivery, 2(2), 859–870, July 1987. Fisher, R.J., Schnetzer, G.H., Thottappillil, R., Rakov, V.A., Uman, M.A., and Goldberg, J.D., Parameters of triggered-lightning flashes in Florida and Alabama, J. Geophys. Res., 98(D12), 22,887–22, 902, 1993. IEEE Guide for Surge Protection of Equipment Connected to AC Power and Communications Circuits, IEEE Press, New York, 2005, Standards Information Network, IEEE Press, p. 3. IEEE Std 1410–1997, IEEE Guide for Improving the Lightning Performance of Electric Power Distribution Lines, IEEE PES, December 1997, Section 5. Kithil, R., Lightning protection codes: Confusion and costs in the USA, in Proceedings of the 24th International Lightning Protection Conference, Birmingham, U.K., September 16, 1998. Leteinturier, C., Weidman, C., and Hamelin, J., Current and electric field derivatives in triggered lightning return strokes, J. Geophys. Res., 95(D1), 811–828, 1990. MacGorman, D.R., Maier, M.W., and Rust, W.D., Lightning strike density for the contiguous United States from thunderstorm duration records, NUREG/CR-3759, Office of Nuclear Regulatory Research, U.S. Nuclear Regulatory Commission, Washington, DC, 44 pp., 1984. Munukutla, K., Vittal, V., Heydt, G., Chipman, D., and Brian Keel, D., A practical evaluation for surge arrester placement for transmission line lightning protection, IEEE Trans. Power Delivery, 25(3), 1742–1748, July 2010. Rakov, M.A., Uman, M.A., and Thottappillil, R., Review of lightning properties from electric field and TV observations, J. Geophys. Res., 99, 10,745–10,750, 1994. Thottappillil, R., Rakov, V.A., Uman, M.A., Beasley, W.H., Master, M.J., and Shelukhin, D.V., Lightning subsequent stroke electric field peak greater than the first stroke and multiple ground terminations, J. Geophys. Res., 97, 7503–7509, 1992. Torres, H., Variations of lightning parameter magnitudes within space and time, in 24th International Conference on Lightning Protection, Birmingham, U.K., September 1998. Uman, M.A. Lightning, Dover, New York, 1969, Appendix E. Uman, M.A., The Lightning Discharge, International Geophysics Series, Vol. 39, Academic Press, Orlando, FL, Chapter 1, 1987. Weidman, C.D. and Krider, E.P., Variations à l’échelle submicroseconde des champs électromagnetiqués rayonnes par la foudre, Ann. Telecommun., 39, 165–174, 1984.

7 Overvoltages Caused by Direct Lightning Strokes

Pritindra Chowdhuri Tennessee Technological University

7.1 7.2

Direct Strokes to Unshielded Lines................................................. 7-2 Direct Strokes to Shielded Lines...................................................... 7-4

7.3 7.4

Significant Parameters...................................................................... 7-9 Outage Rates by Direct Strokes......................................................7-11

Shielding Design

Unshielded Lines • Shielded Lines

7.5 Effects of Induction for Direct Strokes..........................................7-15 7.6 Ground Impedance and Corona under Lightning......................7-16 References......................................................................................................7-17

A lightning stroke is defined as a direct stroke if it hits either the tower or the shield wire or the phase conductor. This is illustrated in Figure 7.1. When the insulator string at a tower flashes over by direct hit either to the tower or to the shield wire along the span, it is called a backflash; if the insulator string flashes over by a strike to the phase conductor, it is called a shielding failure for a line shielded by shield wires. Of course, for an unshielded line, insulator flashover is caused by backflash when the stroke hits the tower or by direct contact with the phase conductor. In the analysis of performance and protection of power systems, the most important parameter that must be known is the insulation strength of the system. It is not a unique number. It varies according to the type of the applied voltage, for example, DC, AC, lightning, or switching surges. For the purpose of lightning performance, the insulation strength has been defined in two ways: basic impulse insulation level (BIL) and critical flashover voltage (CFO or V50). BIL has been defined in two ways. The statistical BIL is the crest value of a standard (1.2/50 μs) lightning impulse voltage, which the insulation will withstand with a probability of 90% under specified conditions. The conventional BIL is the crest value of a standard lightning impulse voltage, which the insulation will withstand for a specific number of applications under specified conditions. CFO or V50 is the crest value of a standard lightning impulse voltage, which the insulation will withstand during 50% of the applications. In this chapter, we will use the conventional BIL as the insulation strength under lightning impulse voltages. Analysis of direct strokes to overhead lines can be divided into two classes: unshielded lines and shielded lines. The first discussion involves the unshielded lines.

7-1

7-2

Power Systems 1

2 3

FIGURE 7.1 Illustration of direct lightning strokes to line: (1) backflash caused by direct stroke to tower, (2) backflash caused by direct stroke to shield wire, and (3) insulator flashover by direct stroke to phase conductor (shielding failure).

7.1 Direct Strokes to Unshielded Lines If lightning hits one of the phase conductors, the return-stroke current splits into two equal halves, each half traveling in either direction of the line. The traveling current waves produce traveling voltage waves, which are given by

V=

Zo I 2

(7.1)

where I is the return-stroke current Zo is the surge impedance of the line given by Z o = (L/C)1/2, and L and C are the series inductance and capacitance to ground per meter length of the line These traveling voltage waves stress the insulator strings from which the line is suspended as these voltages arrive at the succeeding towers. The traveling voltages are attenuated as they travel along the line by ground resistance and mostly by the ensuing corona enveloping the struck line. Therefore, the insulators of the towers adjacent to the struck point are most vulnerable. If the peak value of the voltage, given by Equation 7.1, exceeds the BIL of the insulator, then it might flash over, causing an outage. The minimum return-stroke current that causes an insulator flashover is called the critical current, Ic, of the line for the specified BIL. Thus, following Equation 7.1,

Ic =

2BIL Zo

(7.2)

Lightning may hit one of the towers. The return-stroke current then flows along the struck tower and over the tower-footing resistance before being dissipated in the earth. The estimation of the insulator voltage in that case is not simple, especially because there has been no consensus about the modeling of the tower in estimating the insulator voltage. In the simplest assumption, the tower is neglected. Then, the tower voltage, including the voltage of the cross arm from which the insulator is suspended, is the voltage drop across the tower-footing resistance given by Vtf = IRtf, where Rtf is the tower-footing resistance. Neglecting the power-frequency voltage of the phase conductor, this is then the voltage across the insulator.

7-3

Overvoltages Caused by Direct Lightning Strokes

It should be noted that this voltage will be of opposite polarity to that for stroke to the phase conductor for the same polarity of the return-stroke current. Neglecting the tower may be justified for short towers. The effect of the tower for transmission lines must be included in the estimation of the insulator voltage. For these cases, the tower has also been represented as an inductance. Then the insulator voltage is given by Vins = Vtf + L(dI/dt), where L is the inductance of the tower. However, it is known that voltages and currents do travel along the tower. Therefore, the tower should be modeled as a vertical transmission line with a surge impedance, Zt, where the voltage and current waves travel with a velocity, vt. The tower is terminated at the lower end by the tower-footing resistance, Rtf, and at the upper end by the lightning channel that can be assumed to be another transmission line of surge impedance, Zch. Therefore, the traveling voltage and current waves will be repeatedly reflected at either end of the tower while producing voltage at the cross arm, Vca. The insulator from which the phase conductor is suspended will then be stressed at one end by Vca (to ground) and at the other end by the power-frequency phase-to-ground voltage of the phase conductor. Neglecting the power-frequency voltage, the insulator voltage, Vins, will be equal to the cross-arm voltage, Vca. This is schematically shown in Figure 7.2a. The initial voltage traveling down the tower, Vto, is Vto(t) = Zt I(t), where I(t) is the initial tower current that is a function of time, t. The voltage reflection coefficients at the two ends of the tower are given by ar1 =

Rtf − Z t Rtf + Z t

and ar 2 =

Zch − Z t Zch + Z t

(7.3)

Figure 7.2b shows the lattice diagram of the progress of the multiply reflected voltage waves along the tower. The lattice diagram, first proposed by Bewley (1951), is the space–time diagram that shows the position and direction of motion of every incident and reflected and transmitted wave on the system at every instant of time. In Figure 7.2, if the heights of the tower and the cross arm are ht and hca, respectively, and the velocity of the traveling wave along the tower is vt, then the time of travel from the tower top to its foot is τt = h t /v t, and the time of travel from the cross arm to the tower foot is τca = h ca /v t.

Zch 2τ

0

4τ

6τ

8τ

Time

ar2 Vca

1

2

3

2τca

4

5

6

7

8

9

vt

ht

Zt

hca

ar1 Rtf (a)

(b)

FIGURE 7.2 Lightning channel striking tower top: (a) schematic of struck tower; (b) voltage lattice diagram.

7-4

Power Systems 1.2 Legend Rtf = 10 Ω Rtf = 20 Ω

1

Vins (MV)

0.8 0.6 0.4 0.2 0

1

0

2

3

4

5 6 Time (µs)

7

8

9

10

FIGURE 7.3 Profiles of insulator voltage for an unshielded line for a lightning stroke to tower. Tower height = 30 m; cross-arm height = 27 m; phase-conductor height = 25 m; cross-arm width = 2 m; return-stroke current = 1/50 ms 30 kA; Zt = 100 Ω; Z ch = 500 Ω.

In Figure 7.2b, the two solid horizontal lines represent the positions of the tower top and the tower foot, respectively. The broken horizontal line represents the cross-arm position. It takes (τt − τca) seconds for the traveling wave to reach the cross arm after lightning hits the tower top at t = 0. This is shown by point 1 in Figure 7.2b. Similarly, the first reflected wave from the tower foot (point 2 in Figure 7.2b) reaches the cross arm at t = (τt + τca). The first reflected wave from the tower top (point 3 in Figure 7.2b) reaches the cross arm at t = (3τt − τca). The downward-moving voltage waves will reach the cross arm at t = (2n − 1) τt − τca , and the upward-moving voltage waves will reach the cross arm at t = (2n − 1) τt + τca , where n = 1, 2, …, n. The cross-arm voltage, Vca(t), is then given by n

Vca (t ) =

∑ (a a

r1 r 2

)n −1Vto (t − (2n − 1)τ t + τ ca )u(t − (2n − 1)τ t + τ ca )

n =1

n

+ ar1

∑ (a a

r1 r 2

)n −1Vto (t − (2n − 1)τ t − τ ca )u(t − (2n − 1)τ t − τ ca )

(7.4)

n =1

The voltage profiles of the insulator voltage, Vins(=Vca), for two values of tower-footing resistances, Rtf, are shown in Figure 7.3. It should be noticed that the Vins is higher for higher Rtf and that it approaches the voltage drop across the tower-footing resistance (IRtf ) with time. However, the peak of Vins is significantly higher than the voltage drop across Rtf. Higher peak of Vins will occur for (1) taller tower and (2) shorter front time of the stroke current (Chowdhuri, 2004).

7.2 Direct Strokes to Shielded Lines One or more conductors are strung above and parallel to the phase conductors of single- and doublecircuit overhead power lines to shield the phase conductors from direct lightning strikes. These shield wires are generally directly attached to the towers so that the return-stroke currents are safely led to ground through the tower-footing resistances. Sometimes, the shield wires are insulated from the towers by short insulators to prevent power-frequency circulating currents from flowing in the closed- circuit loop formed by the shield wires, towers, and the earth return. When lightning strikes the shield wire, the short insulator flashes over, connecting the shield wire directly to the grounded towers.

7-5

Overvoltages Caused by Direct Lightning Strokes

For a shielded line, lightning may strike either a phase conductor or the shield wire or the tower. If it strikes a phase conductor but the magnitude of the current is below the critical current level, then no outage occurs. However, if the lightning current is higher than the critical current of the line, then it will precipitate an outage, which is called the shielding failure. In fact, sometimes, shielding is so designed that a few outages are allowed, the objective being to reduce excessive cost of shielding. However, the critical current for a shielded line is higher than that for an unshielded line because the presence of the grounded shield wire reduces the effective surge impedance of the line. The effective surge impedance of a line shielded by one shield wire is given by (Chowdhuri, 2004)

2 Z12 Z22

(7.5)

dp s 2hp 2h ; Z22 = 60n s ; Z12 = 60n ′ dps rp rs

(7.6)

Zeq = Z11 −

Z11 = 60n

where hp and rp are the height and the radius of the phase conductor hs and rs are the height and the radius of the shield wire dp ′s is the distance from the shield wire to the image of the phase conductor in the ground dps is the distance from the shield wire to the phase conductor Z11 is the surge impedance of the phase conductor in the absence of the shield wire, Z22 is the surge impedance of the shield wire, and Z12 is the mutual surge impedance between the phase conductor and the shield wire. It can be shown that either for strokes to tower or for strokes to shield wire, the insulator voltage will be the same if the attenuation caused by impulse corona on the shield wire is neglected (Chowdhuri, 2004). For a stroke to tower, the return-stroke current will be divided into three parts; two parts going to the shield wire in either direction from the tower and the third part to the tower. Thus, lower voltage will be developed along the tower of a shielded line than that for an unshielded line for the same returnstroke current, because lower current will penetrate the tower. This is another advantage of a shield wire. The computation of the cross-arm voltage, Vca, is similar to that for the unshielded line, except for the following modifications in Equations 7.3 and 7.7:

1. The initial tower voltage is equal to IZeq, instead of IZt as for the unshielded line, where Zeq is the impedance as seen from the striking point, that is, Zeq =

0.5 Z s Z t 0.5 Z s + Z t

(7.7)

where Z s = 60ℓn(2hs/rs) is the surge impedance of the shield wire. 2. The traveling voltage wave moving upward along the tower, after being reflected at the tower foot, encounters three parallel branches of impedances, the lightning-channel surge impedance, and the surge impedances of the two halves of the shield wire on either side of the struck tower. Therefore, Zch in Equation 7.3 should be replaced by 0.5Z sZch/(0.5Z s + Zch).

The insulator voltage, Vins, for a shielded line is not equal to Vca, as for the unshielded line. The shield-wire voltage, which is the same as the tower-top voltage, Vtt, induces a voltage on the phase conductor by electromagnetic coupling. The insulator voltage is, then, the difference between Vca and this coupled voltage:

Vins = Vca − kspVtt

(7.8)

7-6

Power Systems

where k sp = Z12/Z22. It can be seen that the electromagnetic coupling with the shield wire reduces the insulator voltage. This is another advantage of the shield wire. To compute Vtt, we go back to Figure 7.2. As the cross arm is moved toward the tower top, τca approaches τt, and naturally, at tower top τca = τt. Then, except the wave 1, the pairs of upward-moving and downward-moving voltages (e.g., 2 and 3, and 4 and 5) arrive at the tower top at the same time. Substituting τca = τt in Equation 7.4, and writing at2 = 1 + ar2, we get Vtt: n

Vtt (t ) = Vtou(t ) + at 2ar1

∑ (a a

r1 r 2

)n −1Vto (t − 2nτ t )u(t − 2nτ t )

(7.9)

n =1

From Equation 7.3 at 2 = 1 + ar 2 =

2Zch Zch + Z t

(7.10)

The coefficient, at2, is called the coefficient of voltage transmission. When lightning strikes the tower, equal voltages (IZeq) travel along the tower as well as along the shield wire in both directions. The voltages on the shield wire are reflected at the subsequent towers and arrive back at the struck tower at different intervals as voltages of opposite polarity (Chowdhuri, 2004). Generally, the reflections from the nearest towers are of any consequence. These reflected voltage waves lower the tower-top voltage. The tower-top voltage remains unaltered until the first reflected waves arrive from the nearest towers. The profiles of the insulator voltage for the same line as in Figure 7.3 but with a shield wire are shown in Figure 7.4. Comparing Figures 7.3 and 7.4, it should be noticed that the insulator voltage is significantly reduced for a shielded line for a stroke to tower. This reduction is possible because (1) a part of the stroke current is diverted to the shield wire, thus reducing the initial tower-top voltage (Vto = It Zt, It < I), and (2) the electromagnetic coupling between the shield wire and the phase conductor induces a voltage on the phase conductor, thus lowering the voltage difference across the insulator (Vins = Vca − kspVtt).

1.2 Legend Rtf = 10 Ω Rtf = 20 Ω

1

Vins (MV)

0.8 0.6 0.4 0.2 0

0

1

2

3

4

5 6 Time (µs)

7

8

9

10

FIGURE 7.4 Profiles of insulator voltage for a shielded line for lightning stroke to tower. Tower height = 30 m; cross-arm height = 27 m; phase-conductor height = 25 m; cross-arm width = 2 m; return-stroke current = 1/50 ms 30 kA; Zt = 100 V; Z ch = 500 V.

7-7

Overvoltages Caused by Direct Lightning Strokes

7.2.1 Shielding Design Striking distance of the lightning stroke plays a crucial role in the design of shielding. Striking distance is defined as the distance through which a descending stepped leader will strike a grounded object. Armstrong and Whitehead (1968) and Brown and Whitehead (1969) proposed a simple relation between the striking distance, rs, and the return-stroke current, I (in kA), of the form rs = aI b (m)

(7.11)

where a and b are constants. The most frequently used value of a is 8 or 10, and that of b is 0.65. Let us suppose that a stepped leader with prospective return-stroke current of Is is descending near a horizontal conductor, P (Figure 7.5a). Its striking distance, rs, will be given by Equation 7.11. It will hit the surface of the earth when it penetrates a plane which is rs meters above the earth. The horizontal conductor will be struck if the leader touches the surface of an imaginary cylinder of radius, rs, with its center at the center of the conductor. The attractive width of the horizontal conductor will be ab in Figure 7.5a. It is given by

ab = 2ω p = 2 rs2 − (rs − hp )2 = 2 hp (2rs − hp ) for rs > hp

(7.12a)

ab = 2ω p = 2rs for rs ≤ hp

(7.12b)

and

rs

rs

rs a

rs

b

c

c΄

a

d S

hp

ωp

P

rs

S΄ P

(a)

b rs

(b) b rs S P

rs

(c)

FIGURE 7.5 Principle of shielding: (a) electrogeometric model, (b) shielding principle, and (c) placement of shield wire for perfect shielding.

7-8

Power Systems

where hp is the height of the conductor. For a multiconductor line with a separation distance, dp, between the outermost conductors, the attractive width will be 2ωp + dp. Now, if a second horizontal conductor, S, is placed near P, the attractive width of S will be cd (Figure 7.5b). If S is intended to completely shield P, then the cylinder around S and the rs-plane above the earth’s surface must completely surround the attractive cylinder around P. However, as Figure 7.5b shows, an unprotected width, db, remains. Stepped leaders falling through db will strike P. If S is repositioned to S′ so that the point d coincides with b, then P is completely shielded by S. The procedure to place the conductor, S, for perfect shielding of P is shown in Figure 7.5c. Knowing the critical current, Ic, from Equation 7.2, the corresponding striking distance, rs, is computed from Equation 7.11. A horizontal straight line is drawn at a distance, rs, above the earth’s surface. An arc of radius, rs, is drawn with P as center, which intersects the rs-line above earth at b. Then, an arc of radius, rs, is drawn with b as center. This arc will go through P. Now, with P as radius another arc is drawn of radius rsp, where rsp is the minimum required distance between the phase conductor and a grounded object. This arc will intersect the first arc at S, which is the position of the shield wire for perfect shielding of P. Figure 7.6 shows the placement of a single shield wire above a three-phase horizontally configured line for shielding. In Figure 7.6a, the attractive cylinders of all three phase conductors are contained within the attractive cylinder of the shield wire and the rs-plane above the earth. However, in Figure 7.6b where the critical current is lower, the single shield wire at S cannot perfectly shield the two outer phase conductors. Raising the shield wire helps in reducing the unprotected width; but, in this case, it cannot completely eliminate shielding failure. As the shield wire is raised, its attractive width increases until the shield-wire height reaches the rs-plane above the earth, where the attractive width is the largest, equal to the diameter of the rs-cylinder of the shield wire. Raising the shield-wire height further will not help. In this case, either the insulation strength of the line should be increased (i.e., the critical current increased) or two shield wires should be used. Figure 7.7 shows the use of two shield wires. In Figure 7.7a, all three phase conductors are completely shielded by the two shield wires. However, for smaller Ic (i.e., smaller rs), part of the attractive cylinder of the middle phase conductor is left exposed (Figure 7.7b). This shows that the middle phase conductor may experience shielding failure even when the outer phase conductors are perfectly shielded. In that case, either the insulation strength of the line should be increased or the height of the shield wires raised or both.

rs rs

rs

rs

rs rs

rs S΄

S S

rs P1

(a)

P2

P1

P3

P2 P3

rs

(b)

FIGURE 7.6 Shielding of three-phase horizontally configured line by single shield wire: (a) perfect shielding and (b) imperfect shielding.

7-9

Overvoltages Caused by Direct Lightning Strokes Unprotected zone

rs

S2 rs

S1 P1 P2

S1

S2 rs

P3

P1

rs

P2

P3 rs

(a)

(b)

FIGURE 7.7 Shielding of three-phase horizontally configured line by two shield wires: (a) perfect shielding and (b) imperfect shielding.

Successful design of shielding depends on the accurate knowledge of the striking distance. The stepped leader parameters (cloud height, charge distribution along the leader) as well as the heights of the tower and the shield wires play prominent roles in determining the striking distances to the shield (tower and shield wires). The leader parameters are stochastic. However, no statistical data on these parameters, except the polarity, are available now. Extensive discussion on the striking distance can be found in Chowdhuri et al. (2007).

7.3 Significant Parameters The most significant parameter in estimating the insulator voltage is the return-stroke current, that is, its peak value, waveshape, and statistical distributions of the amplitude and waveshape. The waveshape of the return-stroke current is generally assumed to be double exponential where the current rapidly rises to its peak exponentially and subsequently decays exponentially: I (t ) = I o (e − a1t − e − a2t )

(7.13)

The parameters, Io, a1, and a2, are determined from the given peak, Ip, the front time, tf, and the time to half value, t h, during its subsequent decay. However, the return-stroke current can also be simulated as a linearly rising and linearly falling wave:

I (t ) = α1tu(t ) − α 2 (t − t f )u(t − t f )

(7.14)

where

α1 =

Ip tf

and α 2 =

2t h − t f Ip 2t f (t h − t f )

(7.15)

Io, a1, and a2 of the double exponential function in Equation 7.13 are not very easy to evaluate. In contrast, α1 and α2 of the linear function in Equation 7.14 are easy to evaluate as given in Equation 7.15. The results from the two waveshapes are not significantly different, particularly for lightning currents

7-10

Power Systems

where tf is on the order of a few microseconds and t h is several tens of microseconds. As t h is very long compared to tf, the influence of t h on the insulator voltage is not significant. Therefore, any convenient number can be assumed for t h (e.g., 50 ms) without the loss of accuracy. The statistical variations of the peak return-stroke current, Ip, fit the log-normal distribution (Popolansky, 1972). The probability density function of Ip, p(Ip), can then be expressed as 1 e 2π I pσ nIp

p(I p ) =

nI p − nI pm −0.5 σ nI p

2

(7.16)

where σℓnlp is the standard deviation of ℓnIp Ipm is the median value of the return-stroke current, Ip The cumulative probability, Pc, that the peak current in any lightning flash will exceed Ip kA can be determined by integrating Equation 7.16 as follows: Substituting, u=

nI p − nI pm

Pc (I p ) =

1 π

(7.17)

du = 0.5 erfc(u)

(7.18)

2σ nIp

∞

∫e

− u2

u

The probability density function, p(tf ), of the front time, tf, can be similarly determined by replacing Ipm and σℓnlp by the corresponding tfm and σℓntf in Equations 7.17 and 7.18. Assuming no correlation between Ip and tf, the joint probability density function of Ip and tf is p(Ip, tf ) = p(Ip)p(tf ). The equation for p(Ip, tf ) becomes more complex if there is a correlation between Ip and tf (Chowdhuri, 2004). The joint probability density function is then given by 2

p(I p , t f ) =

(

(0.5/e1− ρ ) f1 − 2ρ f1 ⋅ f 2 + f 2

(

2π(I p ⋅ t f ) σ nIp ⋅ σ nt f

)

2

1−ρ

)

(7.19)

where 2

nI p − nI pm f1 = ; σ nIp

nt f − nt fm f2 = σ

2

nt f

and ρ = coefficient of correlation. The statistical parameters (Ipm, σℓnIp, tfm, and σℓntf) have been analyzed in Anderson and Eriksson (1980) and Eriksson (2005) and are given in Chowdhuri (2004) and IEEE = PES Task Force 15.09 (2005): t fm = 3.83 µs; σ nt f = 0.553 For I p < 20 kA : I pm = 61.1 kA; σ nIp = 1.33

For I p > 20 kA : I pm = 33.3 kA; σ nIp = 0.605

Overvoltages Caused by Direct Lightning Strokes

7-11

Besides Ip and tf, the ground flash density, ng, is the third significant parameter in estimating the lightning performance of power systems. The ground flash density is defined as the average number of lightning flashes per square kilometer per year in a geographic region. It should be borne in mind that the lightning activity in a particular geographic region varies by a large margin from year to year. Generally, the ground flash density is averaged over 10 years. In the past, the index of lightning severity was the keraunic level, that is, the number of thunder days in a region, because that was the only parameter available. Several empirical equations have been used to relate keraunic level with ng. However, there has been a concerted effort in many parts of the world to measure ng directly, and the measurement accuracy has also been improved in recent years.

7.4 Outage Rates by Direct Strokes The outage rate is the ultimate gauge of lightning performance of a transmission line. It is defined as the number of outages caused by lightning per 100 km of line length per year. One needs to know the attractive area of the line in order to estimate the outage rate. The line is assumed to be struck by lightning if the stroke falls within the attractive area. The electrical shadow method has been used to estimate the attractive area. According to the electrical shadow method, a line of height, hp meters, will attract lightning from a distance of 2 hp m from either side. Therefore, for a 100 km length, the attractive area will be 0.4 hp km2. This area is then a constant for a specific overhead line of given height and is independent of the severity of the lightning stroke (i.e., Ip). The electrical shadow method has been found to be unsatisfactory in estimating the lightning performance of an overhead power line. Now, the electrogeometric model is used in estimating the attractive area of an overhead line. The attractive area is estimated from the striking distance, which is a function of the return-stroke current, Ip, as given by Equation 7.11. Although it has been suggested that the striking distance should also be a function of other variables (Chowdhuri and Kotapalli, 1989), the striking distance as given by Equation 7.11 is being universally used. The first step in the estimation of outage rate is the determination of the critical current. If the returnstroke current is less than the critical current, then the insulator will not flash over if the line is hit by the stepped leader. If one of the phase conductors is struck, such as for an unshielded line, then the critical current is given by Equation 7.2. However, for strikes either to the tower or to the shield wire of a shielded line, the critical current is not that simple to compute if the multiple reflections along the tower are considered as in Equation 7.4 or 7.9. For these cases, it is best first to compute the insulator voltage by Equation 7.4 or 7.9 for a return-stroke current of 1 kA, and then estimate the critical current by taking the ratio between the insulation strength and the insulator voltage caused by 1 kA of return-stroke current of the specified front time, tf, bearing in mind that the insulator voltage is a function of tf. Methods of estimation of the outage rate for unshielded and shielded lines will be somewhat different. Therefore, they are discussed separately.

7.4.1 Unshielded Lines The vertical towers and the horizontal phase conductors coexist for an overhead power line. In that case, there is a race between the towers and the phase conductors to catch the lightning stroke. Some lightning strokes will hit the towers and some will hit the phase conductors. Figure 7.8 illustrates how to estimate the attractive areas of the towers and the phase conductors. The tower and the two outermost phase conductors are shown in Figure 7.8. In the cross-sectional view, a horizontal line is drawn at a height rs from the earth’s surface, where rs is the striking distance corresponding to the return-stroke current, Is. A circle (cross-sectional view of a sphere) is drawn with radius, rs, and center at the tip of the tower, cutting the line above the earth at a and b. Two circles (representing cylinders) are drawn with radius, rs, and centers at the outermost phase conductors, cutting the line above the earth again at a and b. The horizontal distance between the tower tip and either a or b is ωt.

7-12

Power Systems Cylinders around lines rs rs a

Sphere around tower top

rs b

ωt

rs

ωp

ht

hp

dp

Earth

Cross-sectional view Sphere around tower top

rs c

d ω

a

rs

Cylinder around line

rs

b ωt ht

Line hp

Earth

Side view ωt ωp

ωt

dp

ωℓ Tower top

ωp

Plan view

FIGURE 7.8 Attractive areas of tower and horizontal conductors.

The side view of Figure 7.8 shows where the sphere around the tower top penetrates both the rs-plane (a and b) above ground and the cylinders around the outermost phase conductors (c and d). The projection of the sphere around the tower top on the rs-plane is a circle of radius, ωt, given by

ω t = rs2 − (rs − ht )2 = ht (2rs − ht )

(7.20)

The projection of the sphere on the upper surface of the two cylinders around the outer phase conductors will be an ellipse with its minor axis, 2ωℓ, along a line midway between the two outer phase conductors and parallel to their axes; the major axis of the ellipse will be 2ωt, as shown in the plan view of Figure 7.8. ωℓ is given by

ω = rs2 − (rs − ht + hp )2

(7.21)

Overvoltages Caused by Direct Lightning Strokes

7-13

If a lightning stroke with return-stroke current, Is or greater, falls within the ellipse, then it will hit the tower. It will hit one of the phase conductors if it falls outside the ellipse but within the width (2ωp + dp); it will hit the ground if it falls outside the width (2ωp + dp). Therefore, for each span length, ℓs, the attractive areas for the tower (At) and for the phase conductors (Ap) will be

At = πω t ω

(7.22a)

Ap = (2ω p + dp ) s − At

(7.22b)

and

The previous analysis was performed for the shielding current of the overhead line when the sphere around the tower top and the cylinders around the outer phase conductors intersect the rs-plane above ground at the same points (points a and b in Figure 7.8). In this case, 2ωt = 2ωp + dp. The sphere and the cylinders will intersect the rs-plane at different points for different return-stroke currents; their horizontal segments (widths) can be similarly computed. The equation for ωt is given earlier. The equation for ωp is given in Equation 7.12. Due to conductor sag, the effective height of a conductor is lower than that at the tower. The effective height is generally assumed as

2 hp = hpt − (midspan sag) 3

(7.23)

where hpt is the height of the conductor at tower. The critical current, icp, for stroke to a phase conductor is computed from Equation 7.2. It should be noted that icp is independent of the front time, tf, of the return-stroke current. The critical current, ict, for stroke to tower is a function of tf. Therefore, starting with a short tf, such as 0.5 ms, the insulator voltage is determined with 1 kA of tower injected current; then, the critical tower current for the selected tf is determined by the ratio of the insulation strength (e.g., BIL) to the insulator voltage determined with 1 kA of tower injected current. The procedure for estimating the outage rate is started with the lower of the two critical currents (icp or ict). If icp is the lower one, which is usually the case, the attractive areas, Ap and At, are computed for that current. If icp < ict, then this will not cause any flashover if it falls within At. In other words, the towers act like partial shields to the phase conductors. However, all strokes with icp and higher currents falling within Ap will cause flashover. The cumulative probability, Pc(icp), for strokes with currents icp and higher is given by Equation 7.18. If there are nsp spans per 100 km of the line, then the number of outages for lightning strokes falling within Ap along the 100 km stretch of the line will be

nfpo = ng Pc (icp ) p(t f )∆t f nsp Ap

(7.24)

where p(tf ) is the probability density function of tf Δtf is the front step size The stroke current is increased by a small step (e.g., 500 A), Δi (i = icp + Δi), and the enlarged attractive area, Ap1, is calculated. All strokes with currents i and higher falling within Ap1 will cause outages. However, outage rate for strokes falling within Ap for strokes icp and greater has already been computed in Equation 7.24. Therefore, only the additional outage rate, Δnfp, should be added to Equation 7.24:

∆nfp = ng Pc (i) p(t f )∆t f nsp ∆Ap

(7.25)

7-14

Power Systems

where ΔAp = Ap1 − Ap. The stroke current is increased in steps of Δi and the incremental outages are added until the stroke current is very high (e.g., 200 kA) when the probability of occurrence becomes acceptably low. Then, the front time, tf, is increased by a small step, Δtf, and the computations are repeated until the probability of occurrence of higher tf is low (e.g., tf = 10.5 μs). In the mean time, if the stroke current becomes equal to ict, then the outages due to strokes to the tower should be added to the outages caused by strokes to the phase conductors. The total outage rate is then given by (Chowdhuri et al., 2002) nfo = nfp + nft

nfp = nfpo + ngnsp

(7.26a)

∑ ∑ P (i) p(t )∆t ∆A c

tf

f

f

p

(7.26b)

i

and nft = nfto + ngnt

∑ ∑ P (i) p(t )∆t ∆A c

tf

f

f

(7.26c)

t

i

With digital computers, the total outage rates can be computed within a few seconds.

7.4.2 Shielded Lines For strokes to the shield wire, the voltage at the adjacent towers will be the same as that for stroke to the tower for the same stroke current. Therefore, there will be only one critical current for strokes to shielded lines, unlike the unshielded lines. The critical current for shielded lines can be computed similar to that for the unshielded lines, except that Equation 7.8 is now used instead of Equation 7.4. Otherwise, the computation for shielded lines is similar to that for unshielded lines. hp and dp for the phase conductors are replaced by hs and ds, which are the shield-wire height and the separation distance between the shield wires, respectively. For a line with a single shield wire, ds = 0. Generally, shield wires are attached to the tower at its top. However, the effective height of the shield wire is lower than that of the tower due to sag. The effective height of the shield wire, hs, can be computed from Equation 7.23 by replacing hpt by hst, the shield-wire height at tower. The backflashover rates of a horizontally configured three-phase overhead line with two shield wires are computed. The dimensions of the line are shown in Table 7.1. The backflashover rates with and without shield wires are plotted in Figure 7.9 as function of the insulation withstand voltage. It should be noted that the outage rate of the unshielded line decreases much slower than that for the shielded line. This happens because the phase conductor is struck by lightning more often than the towers, and the critical current for strikes to the phase conductor is significantly smaller than that for strike to a tower. Tower surge impedance = 100 Ω; tower-footing resistance = 10 Ω; return-stroke surge impedance = 500 Ω; ground flash density = 10/km2/year; correlation coefficient = 0.47. TABLE 7.1 Dimensions of a Horizontally Configured Distribution Line Cross-Arm Height (m) 9.72

Phase-Conductor Height (m)

Phase-Conductor Separation (m)

Shield-Wire Height (m)

Shield-Wire Separation (m)

Span Length (m)

10.00

3.66

13.05

3.66

100.0

7-15

Overvoltages Caused by Direct Lightning Strokes 80

No. of outages/100 km/year

70 60 50 40 30 20 Unshielded Shielded

10 0

50

100

150

200

250

300

350

400

450

500

550

600

Insulation withstand voltage (kV)

FIGURE 7.9 Outage rates of the horizontally configured three-phase overhead line of Table 7.1 (shielded and unshielded) caused by direct strokes to the line.

7.5 Effects of Induction for Direct Strokes In the analysis of the insulator-string voltage, Vins, it was assumed that Vins has two components: (1) one due to the cross-arm voltage, Vca, caused by the multiply reflected traveling voltage waves along the struck tower and (2) the other due to electromagnetic coupling of the phase conductor to the shield wire for a shielded line. However, two other component voltages contribute to the total voltage across the insulator string (Chowdhuri et al., 2002; Chowdhuri, 2004). The third voltage component is induced by the electromagnetic fields of the charge and current of the return-stroke channel (Figure 7.10). The stroke channel being only a few meters from the phase conductors, the induction effect should not be ignored. The fourth voltage component is induced across the insulator string by the multiply reflected traveling current waves along the struck tower (Figure 7.10). The voltage induced

Lightning channel

Shield wire Phase conductor Tower Footing resistance Earth

FIGURE 7.10 Direct stroke to a tower.

7-16

Power Systems 90

No. of outages/100 km/year

80 70 60

Two voltage components Four voltage components

50 40 30 20 10 0 50

100

150

200

250

300

350

400

450

500

550

600

Insulation withstand voltage (kV)

FIGURE 7.11 Outage rates by direct strokes of the horizontally configured three-phase overhead line of Table 7.1 considering two and four voltage components across insulator string. Tower surge impedance = 100 Ω; towerfooting resistance = 10 Ω; velocity of traveling-wave current in tower = 0.8 p.u.; cloud height = 3 km; return-stroke surge impedance = 500 Ω; ground flash density = 10/km2/year; correlation coefficient = 0.47.

across the insulator string by the rapid change of the tower current can be significant because of the proximity of the tower to the insulators. The third voltage component can be computed by the analysis shown in Chapter 8, with the difference that, in the present case, the stroke hits the tower top instead of the ground. This difference is manifested in the inducing voltage, Vi, which is given in Equations 8.2 and 8.3. The scalar potential, φ, and the vector potential, A, in Equation 8.3 can be computed from Chapter 8 of Chowdhuri (2004) with the difference that the lower and the upper limits of z′ (height of a specified point along the stepped leader) are the tower height, ht, and the cloud height, hc, for strike to the tower top. Therefore, the voltage induced on the phase conductor for a tower strike is computed for two cloud heights (hc and ht), and then the second induced voltage (for ht) is subtracted from the first induced voltage (due to hc). The fourth voltage component, that is, the induced voltage due to the traveling current along the tower, can be computed similar to the normal case for stroke to ground, with the following modifications:

1. In the computation of the inducing voltage for a strike to ground, both the electrostatic component of the upper charge column and the magnetic component of the lower current column are considered. For the traveling current waves along the tower, only magnetic component of the electromagnetic field is considered. 2. After each reflection at either end of the struck tower, both the magnitude and the direction of travel change for the traveling current waves. Therefore, new computations must be superimposed on the initial computation after each reflection.

The effects of the third and fourth voltage components on the estimation of outage rates are shown in Figure 7.11 for a horizontally configured three-phase distribution line with two shield wires.

7.6 Ground Impedance and Corona under Lightning The lightning current flowing through the overhead lines eventually finds its way to ground through the ground electrodes to the soil. The impedance of the ground electrodes can be computed as discussed in Chisholm (2007). The electrical characteristics of the soil in which the ground electrodes are buried

Overvoltages Caused by Direct Lightning Strokes

7-17

play a significant role in the overall grounding impedance. The electrical resistivity of the soil bed varies from day to day, season to season, and year to year, depending upon the amount of rainfall and temperature. The soil is ionized under large lightning current. The ionization becomes more intense with larger currents. The critical electric field for soil ionization varies for the same soil, depending on its compactness, moisture content and temperature, and varying soil electrical conductivity and permittivity (Manna and Chowdhuri, 2007). All these need to be known to estimate the grounding impedance under lightning currents. Many of these parameters are not known. However, research is going on.

References Anderson, R.B. and Eriksson, A.J., Lightning parameters for engineering applications, Electra, 69, 65–102, 1980. Armstrong, H.R. and Whitehead, E.R., Field and analytical studies of transmission line shielding, IEEE Trans. Power Appar. Syst., PAS-87, 270–281, 1968. Bewley, L.V., Traveling Waves on Transmission Systems, 2nd edn., John Wiley, New York, 1951. Brown, G.W. and Whitehead, E.R., Field and analytical studies of transmission line shielding: Part II, IEEE Trans. Power Appar. Syst., PAS-88, 617–626, 1969. Chisholm, W.A., Transmission system transients—Grounding, power systems, Electric Power Engineering Handbook, 2nd edn., L.L. Grigsby, (ed.), CRC Press, Boca Raton, FL, 2007. Chowdhuri, P., Electromagnetic Transients in Power Systems, 2nd edn., Research Studies Press, Baldock, Hertfordshire, U.K., 2004. Chowdhuri, P. and Kotapalli, A.K., Significant parameters in estimating the striking distance of lightning strokes to overhead lines, IEEE Trans. Power Delivery, 4, 1970–1981, 1989. Chowdhuri, P., Li, S., and Yan, P., Rigorous analysis of back-flashover outages caused by direct lightning strokes to overhead power lines, IEE Proc.—Gener. Transm. Distrib., 149, 58–65, 2002. Chowdhuri, P., Tajali, G.R., and Yuan, X., Analysis of striking distances of lightning strokes to vertical towers, IET Gener. Transm. Distrib., 1, 879–886, 2007. Ericsson, A.J., Notes on lightning parameters, CIGRE Note 33–86 (WG33-01) IWD, 15 July 1986. IEEE = PES Task Force 15.09, Parameters of lightning strokes: A review, IEEE Trans. Power Delivery, 20, 346–358, 2005. Manna, T. K. and Chowdhuri, P., Generalised equation of soil critical electric field Ec based on impulse tests and measured soil electrical parameters, IET Gener. Transm. Distrib., 1, 811–817, 2007. Popolansky, F., Frequency distribution of amplitudes of lightning currents, Electra, 22, 139–147, 1972.

8 Overvoltages Caused by Indirect Lightning Strokes 8.1 8.2 8.3 8.4

Tennessee Technological University

Evaluation of Green’s Function • Induced Voltage Caused by Return-Stroke Current of Arbitrary Waveshape

Induced Voltages on Multiconductor Lines................................ 8-10 Effects of Shield Wires on Induced Voltages............................... 8-12 Stochastic Characteristics of Lightning Strokes......................... 8-13 Estimation of Outage Rates Caused by Nearby Lightning Strokes............................................................................. 8-14 8.9 Estimation of Total Outage Rates.................................................. 8-17 8.A Appendix A: Voltage Induced by Linearly Rising and Falling Return-Stroke Current.............................................. 8-17 References..................................................................................................... 8-18 8.5 8.6 8.7 8.8

Pritindra Chowdhuri

Inducing Voltage................................................................................ 8-3 Induced Voltage.................................................................................8-4 Green’s Function................................................................................ 8-5 Induced Voltage of a Doubly Infinite Single-Conductor Line............................................................... 8-6

A direct stroke is defined as a lightning stroke when it hits a shield wire, a tower, or a phase conductor. An insulator string is stressed by very high voltages caused by a direct stroke. An insulator string can also be stressed by high transient voltages when a lightning stroke hits the nearby ground. An indirect stroke is illustrated in Figure 8.1. During the early years of power transmission, protection against a direct lightning stroke was thought to be impossible. Therefore, power lines were designed with overhead ground wires to protect against transient voltages induced from nearby lightning strokes to ground. As the transmission voltages became higher and higher with the consequent increase in the insulation levels of overhead lines, it was realized that the transient voltages induced from nearby lightning strokes were not the chief source of disturbance, and that protection against direct lightning strokes is indeed technically feasible and economically justifiable (Fortescue, 1930). Now emphasis is put on protection against both direct and indirect lightning strokes. As the reliability of distribution power lines has become increasingly important, many researchers have been studying the effects of lightning-induced voltages on power distribution lines. These various researches have been reviewed by Chowdhuri et al. (2001). The voltage induced on a line by an indirect lightning stroke has four components:

1. The charged cloud above the line induces bound charges on the line while the line itself is held electrostatically at ground potential by the neutrals of connected transformers and by leakage over the insulators. When the cloud is partially or fully discharged, these bound charges are released and travel in both directions on the line giving rise to the traveling voltage and current waves. 8-1

8-2

Power Systems

ld

wir

e

a Ph

se

Indirect stroke

ie Sh

r cto du n co

FIGURE 8.1 Illustration of direct and indirect lightning strokes.

2. The charges lowered by the stepped leader further induce charges on the line. When the stepped leader is neutralized by the return stroke, the bound charges on the line are released and thus produce traveling waves similar to that caused by the cloud discharge. 3. The residual charges on the upper part of the return stroke induce an electrostatic field in the vicinity of the line and hence an induced voltage on it. 4. The rate of change of current in the return stroke produces a magnetically induced voltage on the line.

If the lightning has subsequent strokes, then the subsequent components of the induced voltage will be similar to one or the other of the four components discussed earlier. The magnitudes of the voltages induced by the release of the charges bound either by the cloud or by the stepped leader are small compared with the voltages induced by the return stroke. Therefore, only the electrostatic and the magnetic components induced by the return stroke are considered in the following analysis. The initial computations are performed with the assumption that the charge distribution along the leader stroke is uniform, and that the return-stroke current is rectangular. However, the result with the rectangular current wave can be transformed to that with currents of any other waveshape by the convolution integral (Duhamel’s theorem). It was also assumed that the stroke is vertical and that the overhead line is loss free and the earth is perfectly conducting. The vertical channel of the return stroke is shown in Figure 8.2, where the upper part consists of a column of residual charge that is neutralized by the rapid upward movement of the return-stroke current in the lower part of the channel. Figure 8.3 shows a rectangular system of coordinates where the origin of the system is the point where lightning strikes the surface of the earth. The line conductor is located at a distance yo meters from the origin, having a mean height of hp meters above ground and running along the x-direction. The origin of time (t = 0) is assumed to be the instant when the return stroke starts at the earth level. Cloud

Residual charge I A

Return stroke Earth

FIGURE 8.2 Return stroke with residual charge column.

8-3

Overvoltages Caused by Indirect Lightning Strokes

r΄

r΄

r to uc d n co ne h i L y

r–

Stroke

z

r

x

yo

Stroke

Stroke Line conductor

h yo

x

FIGURE 8.3 Coordinate system of line conductor and lightning stroke.

8.1 Inducing Voltage The total electric field created by the charge and the current in the lightning stroke at any point in space is Ei = Eei + Emi = −∇ϕ −

∂A ∂t

(8.1)

where φ is the inducing scalar potential created by the residual charge at the upper part of the return stroke and A is the inducing vector potential created by the upward-moving return-stroke current (Figure 8.2). φ and A are called the retarded potentials, because these potentials at a given point in space and time are determined by the charge and current at the source (i.e., the lightning channel) at an earlier time; the difference in time (i.e., the retardation) is the time required to travel the distance between the source and the field point in space with a finite velocity, which in air is c = 3 × 108 m/s. These electromagnetic potentials can be deduced from the distribution of the charge and the current in the return-stroke channel. The next step is to find the inducing electric field (Equation 8.1). The inducing voltage, Vi, is the line integral of Ei: hp

∫

hp

∫

hp

∫

Vi = − Ei ⋅ dz = − Eei ⋅ dz − Emi ⋅ dz = Vei + Vmi

0

0

(8.2)

0

As the height, hp, of the line conductor is small compared with the length of the lightning channel, the inducing electric field below the line conductor can be assumed to be constant and equal to that on the ground surface:

∂A Vi = ∇ϕ + ⋅ hp ∂t

(8.3)

The inducing voltage will act on each point along the length of the overhead line. However, because of the retardation effect, the earliest time, to, the disturbance from the lightning channel will reach a point on the line conductor would be

to =

x 2 + y o2 c

(8.4)

8-4

Power Systems

Therefore, the inducing voltage at a point on the line remains zero until t = to. Hence, Vi = ψ(x , t )u(t − t o )

(8.5)

where u(t – to) is the shifted unit step function. The continuous function, ψ(x,t), is the same as Equation 8.3, and is given, for a negative stroke with uniform charge density along its length, by (Rusck, 1958) ψ(x , t ) = −

60I ohp 1 − β2 1 − 2 β β2c 2 (t − t o )2 + (1 − β2 )r 2 hc + r 2

(8.6)

where Io is the step-function return-stroke current (A) hp is the height of line above ground (m) β is the v/c, v is the velocity of return stroke (m/s) r is the distance of point x on line from point of strike (m) hc is the height of cloud charge center above ground (m) The inducing voltage is the voltage at a field point in space with the same coordinates as a corresponding point on the line conductor, but without the presence of the line conductor. The inducing voltage at different points along the length of the line conductor will be different. In the overhead line, these differences will tend to be equalized by the flow of current, as it is a good conductor of electricity. Therefore, the actual voltage between a point on the line and the ground below it will be different from the inducing voltage at that point. This voltage, which can actually be measured on the line conductor, is defined as the induced voltage. The calculation of the induced voltage is the primary objective.

8.2 Induced Voltage Neglecting losses, an overhead line may be represented as consisting of distributed series inductance L (H/m) and distributed shunt capacitance C (F/m). The effect of the inducing voltage will then be equivalent to connecting a voltage source along each point of the line (Figure 8.4). The partial differential equation for such a configuration will be −

I

∂V ∂I ∆x = L∆x ∂x ∂t ∂I

I+

∂x

(8.7)

∆x

L ∆x

V

C ∆x Vi +

Vi

∂Vi ∂x

∆x

FIGURE 8.4 Equivalent circuit of transmission line with inducing voltage.

∆x

Overvoltages Caused by Indirect Lightning Strokes

8-5

and −

∂I ∂ ∆x = C ∆x (V − Vi ) ∂x ∂t

(8.8)

Differentiating Equation 8.7 with respect to x, and eliminating I, the equation for the induced voltage can be written as ∂2V 1 ∂2V 1 ∂2V − 2 2 = − 2 2i = F ( x , t ) 2 ∂x c ∂t c ∂t

where

c=

1 = LC

1 = 3 × 108 m/s µoεo

(8.9)

(8.10)

In Laplace transform,

s2 ∂2V (x , s) s 2 V ( x , s ) Vi (x , s) = F (x , s) − = − ∂x 2 c2 c2

(8.11)

Equation 8.9 is an inhomogeneous wave equation for the induced voltage along the overhead line. It is valid for any charge distribution along the leader channel and any waveshape of the return-stroke current. Its solution can be obtained by assuming F(x,t) to be the superposition of impulses that involves the definition of Green’s function (Morse and Feshbach, 1950).

8.3 Green’s Function To obtain the voltage caused by a distributed source, F(x), the effects of each elementary portion of the source are calculated and then integrated for the whole source. If G(x; x′) is the voltage at a point x along the line caused by a unit impulse source at a source point x′, the voltage at x caused by a source distribution F(x′) is the integral of G(x; x′)F(x′) over the whole domain (a,b) of x′ occupied by the source, provided that F(x′) is a piecewise continuous function in the domain a ≤ x′ ≤ b: b

∫

V (x ) = G(x ; x ′)F (x ′) dx ′

(8.12)

a

The function G(x; x′), called Green’s function, is, therefore, a solution for a case that is homogeneous everywhere except at one point. Green’s function, G(x; x′), has the following properties:

1. G(x ; x ′ + 0) − G(x ; x ′ − 0) = 0

(8.13)

dG dG − = 1 2. dx x ′ + 0 dx x ′ − 0

(8.14)

8-6

Power Systems

3. G(x;x′) satisfies the homogeneous equation everywhere in the domain, except at the point x = x′ 4. G(x;x′) satisfies the prescribed homogeneous boundary conditions.

Green’s function can be found by converting Equation 8.11 to a homogeneous equation and replacing V(x, s) by G(x; x′,s): ∂2G(x ; x ′, s) s 2 − 2 G(x ; x ′, s) = 0 ∂x 2 c

(8.15)

The general solution of Equation 8.15 is given by G(x ; x ′, s) = Ae sx/c + Be − sx/c

(8.16)

The constants A and B are found from the boundary conditions and from the properties of Green’s function.

8.4 Induced Voltage of a Doubly Infinite Single-Conductor Line The induced voltage at any point, x, on the line can be determined by invoking Equation 8.12, where G(x; x′) F(x′) is the integrand. F(x′) is a function of the amplitude and waveshape of the inducing voltage, Vi (Equation 8.5), whereas Green’s function, G(x; x′), is dependent on the boundary conditions of the line and the properties of Green’s function. In other words, it is a function of the line configuration and is independent of the lightning characteristics. Therefore, it is appropriate to determine Green’s function first.

8.4.1 Evaluation of Green’s Function As Green’s function is finite for x → −∞ and x → +∞,

G1 = Ae

sx/c

for x < x ′; G2 = Be − sx/c

for x > x ′

From Equation 8.13,

Ae

sx ′/c

= Be

− ( sx ′/c )

, i.e., B = Ae

2 sx ′/c

From Equation 8.14, A=

c − sx/c e 2s

Hence, B=−

G1(x ; x ′, s) = −

c sx/c e 2s

s(x ′ − x ) c exp − c 2s

for x < x ′

(8.17)

8-7

Overvoltages Caused by Indirect Lightning Strokes

and G2 (x ; x ′, s) = −

c s(x ′ − x ) exp 2s c

for x > x ′

(8.18)

By applying Equation 8.12, V ( x , s) = −

c 2s

∞

x

∫

e s( x ′ − x )/c F (x ′, s) dx ′ −

−∞

c e − s( x ′ − x )/c F (x ′, s) dx ′ = V1(x , s) + V2 (x , s) 2s

∫

(8.19)

x

8.4.2 Induced Voltage Caused by Return-Stroke Current of Arbitrary Waveshape The induced voltage caused by return-stroke current, I(t), of arbitrary waveshape can be computed from Equation 8.11 by several methods. In method I, the inducing voltage, Vi, due to I(t) is found by applying Duhamel’s integral (Haldar and Liew, 1988): t

Vi =

d I (t − τ)Vistep (x ′, τ) dτ dt

∫

(8.20)

0

where Vistep is the inducing voltage caused by a unit step-function current. In other words, Vistep (x ′, τ) = ψ o (x ′, τ) u(τ − t o )

(8.21)

where ψ o ( x ′ , τ) =

ψ ( x ′ , τ) Io

and ψ(x′,τ) is given in Equation 8.6. Inserting Equation 8.21 in Equation 8.20 and taking Laplace transform of Vi in Equation 8.20, Vi (x ′, s) = sI (s)ψ o (x ′, s)e − sto

(8.22)

and F ( x ′, s) = −

s2 s2 V ( x , s ) = − I (s)ψ o (x ′, s)e − sto ′ i c2 c2

(8.23)

Replacing F(x′,s) in Equation 8.19 by Equation 8.23, the induced voltage, V(x,s), is

V ( x , s) =

∞ x′−x x′−x x − s to + − s to − 1 c c sI (s) s ψ o (x ′, s)e dx ′ + s ψ o (x ′, s)e dx ′ 2c x −∞

∫

∫

(8.24)

8-8

Power Systems

Inverting to time domain by convolution integral, 1 V (x , t ) = 2c +

t

∫ 0

1 2c

d x x′ − x x′ − x d I (t − τ) ψ o x ′, τ + u τ − to + dx ′ dτ c c dt dτ −∞

∫

t

∫ 0

d ∞ x′ − x x′ − x d dτ = V1(x , t ) + V2 (x , t ) (8.25) I (t − τ) ψ o x ′, τ − u τ − t − d x ′ o dt c c dτ x

∫

Because of the shifted unit step function in V1(x,t), τ ≥ to −

x′ − x c

In the limit, τ = to −

x o1 − x = c

x o21 + y o2 x o1 − x − c c

or x o1 =

y o2 − (cτ − x )2 2(cτ − x )

(8.26)

Similarly, for V2 (x , t ) : x o2 =

(cτ + x )2 − y o2 2(cτ + x )

(8.27)

Replacing −∞ by xo1 in V1(x,t) and +∞ by xo2 in V2(x,t) in Equation 8.25, V1(x , t ) =

1 2c

t

∫ 0

x x′ − x d d I (t − τ) ψ o x , τ + x d ′ u(τ − t o )dτ c dt dτ x o 1

(8.28)

xo 2 d d x′ − x d I (t − τ) ψ o x , τ − x ′ u(τ − t o ) dτ dt dτ c x

(8.29)

∫

and

1 V2 (x , t ) = 2c

t

∫ 0

∫

A lightning return-stroke current can be represented by a linearly rising and linearly falling wave with sufficient accuracy (Figure 8.5) (Chowdhuri, 2004):

I (t ) = α1tu(t ) − α 2 (t − t f )u(t − t f )

(8.30)

where

α1 =

Ip tf

and α 2 =

2t H − t f Ip 2t f (t H − t f )

(8.31)

8-9

Overvoltages Caused by Indirect Lightning Strokes

I1 (t) Ip

I (t) = I1(t) + I2(t)

0.5 Ip

tH

tf

t

I 2(t)

FIGURE 8.5 A linearly rising and falling lightning return-stroke current.

It will be evident from Equation 8.30 that V1(x,t) in Equation 8.28 will have two components: one component, V11(x,t), will be a function of I1(t), and the other component, V21(x,t), will be a function of I2(t), that is, V1(x,t) = V11(x,t) + V21(x,t). Similarly, V2(x,t) = V12(x,t) + V22(x,t). After integration and simplifying Equation 8.28, V11(x,t) can be written as

V11(x , t ) = −

α1hp f (τ = t ) f (τ = t ) f 21(τ = t o ) + n 31 × 10 −7 u(t − t o ) (1 − β2 )n 11 β τ t ) f ( τ = t ) f f ( = o 21 31 (τ = t o ) 11

(8.32)

where 2 2 f11(τ) = m11 + m11 + a11 ;

2 2 f 21(τ) = m21 + m21 + a11 ;

f 31 = x o1 + x o21 + y o2 + hc2 ;

2 m11 = x + β2 (cτ − x ); m21 = x o1 + β2 (cτ − x ); a11 = (1 − β2 )[ y o2 + β2 (cτ − x)2 ]

The expression for V21(x,t) is similar to Equation 8.32, except that α1 is replaced by (−α2), and t is replaced by (t − tf ). The computation of V2(x,t) is similar; namely,

V12 (x , t ) = −

α1hp f (τ = t ) f (τ = t ) f 22 (τ = t o ) − n 32 × 10 −7 u(t − t o ) (1 − β2 )n 12 β τ t ) f ( τ = t ) f f ( = o 22 32 (τ = t o ) 12

(8.33)

where 2 2 f12 (τ) = m12 + m12 + a12 ;

2 2 f 22 = m22 + m22 + a12 ;

f 32 = x o2 + x o22 + y o2 + hc2 ;

2 m12 = x o2 − β2 (cτ + x ); m22 = x − β2 (cτ + x ); a12 = (1 − β2 )[ y o2 + β2 (cτ + x )2 ]

V22(x,t) can be similarly determined by replacing α1 in Equation 8.33 by (−α2) and replacing t by (t − tf ). The second method of determining the induced voltage, V(x,t), is to solve Equation 8.19, for a unit step-function return-stroke current, and then find the induced voltage for the given return-stroke current waveshape by applying Duhamel’s integral (Chowdhuri and Gross, 1967; Chowdhuri, 1989a). The solution of Equation 8.19 for a unit step-function return-stroke current is given by (Chowdhuri, 1989a):

Vstep (x , t ) = (V11 + V12 + V21 + V22 )u(t − t o )

(8.34)

8-10

Power Systems

where V11 =

30hp (1 − β2 ) (ct − x )x − y o2 ct x β ( − ) + 2 2 2 2 2 2 β2 (ct − x )2 + y o2 c t + (( 1 − β ) / β )( x + y o)

V12 = − V21 =

30hp (1 − β2 ) β (ct + x )2 + y o2

V22 = −

1 30hp 1 1 − − β2 2 ct − x β k + 1 1

2

(ct + x )x + y o2 β(ct + x ) − 2 2 2 2 2 2 c t + (( 1 − β ) / β )( x + y o)

1 30hp 1 1 − − β2 β ct + x k22 + 1

(8.35)

(8.36)

(8.37)

(8.38)

k1 =

2hc (ct − x ) y o2 + (ct − x )2

(8.39)

k2 =

2hc (ct + x ) y o2 + (ct + x )2

(8.40)

and

The expressions for the induced voltage, caused by a linearly rising and falling return-stroke current, are given in Appendix 8.A. The advantage of method II is that once the induced voltage caused by a step-function return-stroke current is computed, it can be used as a reference in computing the induced voltage caused by currents of any given waveshape by applying Duhamel’s integral, thus avoiding the mathematical manipulations for every given waveshape. However, the mathematical procedures are simpler for method I than that for method II. A third method to solve Equation 8.9 is to apply numerical method, which bypasses all mathematical complexities (Agrawal et al., 1980). However, the accuracy of the numerical method strongly depends upon the step size of computation. Therefore, the computation of the induced voltage of long lines, greater than 1 km, becomes impractical.

8.5 Induced Voltages on Multiconductor Lines Overhead power lines are usually three-phase lines. Sometimes several three-phase circuits are strung from the same tower. Shield wires and neutral conductors are part of the multiconductor system. The various conductors in a multiconductor system interact with each other in the induction process for lightning strokes to nearby ground. The equivalent circuit of a two-conductor system is shown in Figure 8.6. Extending to an n-conductor system, the partial differential equation for the induced voltage, in matrix form, is (Chowdhuri and Gross, 1969; Cinieri and Fumi, 1979; Chowdhuri, 1990, 2004):

∂2[V ] 1 ∂2[V ] ∂2[Vi ] ∂2[Vi ] − = − [ ][ ] = − [ ] L C M g ∂x 2 ∂t 2 ∂t 2 c 2 ∂t 2

(8.41)

8-11

Overvoltages Caused by Indirect Lightning Strokes

Vi2 +

Vi2 C2g ∆x

V1 C1g ∆x

∂x

∆x

L2 ∆x

V2 I1

∂Vi2

I12

L1 ∆x

C1–2 ∆x

Vi1

Vi1 +

I11

∂Vi1 ∂x

∆x

FIGURE 8.6 Equivalent circuit of a two-conductor system.

where [L] is an n × n matrix whose elements are

Lrr = 2 × 10 −7 n

2hr d ; Lrs = 2 × 10 −7 n r ′s rr drs

[Cg] is an n × n diagonal matrix whose elements are Cjg = Cj1 + Cj2 + + Cjn, where Cjr is an element of an n × n matrix, [C] = [p]−1 and

prr = 18 × 109 n

2hr d ; prs = 18 × 109 n r ′s rr drs

where hr and rr are the height above ground and the radius of the rth conductor dr′s is the distance between the image of the rth conductor below earth and the sth conductor drs is the distance between the rth and sth conductors From Equation 8.41, for the jth conductor,

∂2V j 1 ∂2V j ∂2V j1 ∂2V jj ∂2V jn M M M − = − + + + + jj jn j1 ∂t 2 ∂x 2 c 2 ∂t 2 ∂t 2 ∂t 2

(8.42)

If the ratio of the inducing voltage of the mth conductor to that of the jth conductor is kmj (m = 1, 2, …, n), then

2 ∂2V j 1 ∂2V j 1 ∂ Vij 2 c M k M M k ( ) − − + + + + = (c 2 M j )Fj (x , t ) 1 1 j j jj jn nj c 2 ∂t 2 ∂x 2 c 2 ∂t 2

(8.43)

where

M j = M j1k1 j + + M jj + + M jnknj and Fj (x , t ) =

2 1 ∂ Vij −c 2 ∂t 2

(8.44)

8-12

Power Systems

If the jth conductor in its present position existed alone, the partial differential equation of its induced voltage, Vjs, would be the same as Equation 8.9, that is,

∂2V js 1 ∂2V js − = Fj (x , t ) ∂x 2 c 2 ∂t 2

(8.45)

Therefore, the ratio of the induced voltage of the jth conductor in an n-conductor system to that of a single conductor at the same position would be

Vj = M jc 2 V js

(8.46)

The inducing voltage being nearly proportional to the conductor height and the lateral distance of the stroke point being significantly larger than the separation distance between phase conductors, the presence of other conductors in a horizontally configured line will be minimal. On the other hand, for a vertically configured line, the induced voltage of the highest conductor will be lower than that for the same conductor without any neighboring conductors. Similarly, the lowest conductor voltage will be pulled up by the presence of the neighboring conductors of higher elevation, and the middle conductor will be the least affected by the presence of the other conductors (Chowdhuri, 2004).

8.6 Effects of Shield Wires on Induced Voltages If there are (n + r) conductors, of which r conductors are grounded (r shield wires), then the partial differential equation for the induced voltages of the n number of phase conductors is given by Chowdhuri and Gross (1969), Cinieri and Fumi (1979), and Chowdhuri (1990, 2004):

∂2[Vn ] 1 ∂2[Vn ] ∂2[Vin ] ∂2[Vin ] − 2 = −[L ′][C gn ] = −[ M g ] 2 2 2 ∂x ∂t ∂t 2 c ∂t

(8.47)

The matrix [L′] is obtained by partitioning the (n + r) × (n + r) inductance matrix of the (n + r) conductors, and putting [L′] = [Lnn] − [Lnr][Lrr]−1[Lrn], where

Lnn [L](n + r ),(n + r ) = Lrn

Lnr Lrr

(8.48)

[Cgn] is an n × n diagonal matrix, each element of which is the sum of the elements of the corresponding row, up to the nth row, of the original (n + r) × (n + r) capacitance matrix of the (n + r) conductors, [C] = [p]−1, where [p] is the matrix of the potential coefficients of the (n + r) conductors. The jth element of [Cgn] is given by n+r

C jgn ( j ≤ n) =

∑C

jk

(8.49)

k =1

From Equation 8.47, the induced voltage of the j-th conductor is

2 ∂2V j 1 ∂2V j 1 ∂ Vij 2 c ( M k M M k ) = −(c 2 M gi )Fj (x , t ) − = − + + + + 1 1 gj j gjj gjn nj c 2 ∂t 2 ∂x 2 c 2 ∂t 2

(8.50)

8-13

Overvoltages Caused by Indirect Lightning Strokes

Defining the protective ratio as the ratio of the induced voltages on the jth conductor with and without the shield wires in place, Protective ratio =

M gj Mj

(8.51)

where Mj is given by Equation 8.44.

8.7 Stochastic Characteristics of Lightning Strokes The voltage induced on an overhead line is caused by the interaction between various lightning returnstroke parameters and the parameters of the line. The most important return-stroke parameters are (a) peak current, Ip, (b) current front time, tf, (c) return-stroke velocity, v, and (d) ground flash density, ng. These parameters are stochastic in nature. Analysis of field data shows that the statistical variation of the peak, Ip, and the time to crest, tf, of the return-stroke current fit lognormal distribution (Anderson and Eriksson, 1980). The probability density function, p(Ip), of Ip then can be expressed as p(I p ) =

e −0.5 f1 I p ⋅ σ nIp ⋅ 2π

(8.52)

where 2

nI p − nI pm f1 = σ nIp

(8.53)

and σℓnIp is the standard deviation of ℓnIp, and Ipm is the median value of Ip. Similarly, the probability density function of tf can be expressed as p(t f ) =

e −0.5 f2 t f ⋅ σ nt f ⋅ 2π

(8.54)

where 2

nt f − nt fm f2 = ∂

(8.55)

nt f

The joint probability density function, p(Ip,tf ), is given by p(I p , t f )=

e(0.5/(1 − ρ2 ))( f1 − 2ρ f1 ⋅ f 2 + f 2 ) 2π(I p ⋅ t f )(σ nIp ⋅ σ nt f ) 1 − ρ2

(8.56)

where ρ is the coefficient of correlation. The statistical parameters of return-stroke current are as follows (Anderson and Eriksson, 1980; Eriksson, 1986; IEEE = PES Task Force 15.09, 2005): Median time to crest, tfm = 3.83 μs; Log (to base e) of standard deviation, σ(ℓntf ) = 0.553. For Ip < 20 kA: Median peak current, Ipm1 = 61.1 kA; Log (to base e) of standard deviation, σ(ℓnIp1) = 1.33. For Ip > 20 kA: Median peak current, Ipm2 = 33.3 kA; Log (to base e) of standard deviation, σ(ℓnIp2) = 0.605. Correlation coefficient, ρ = 0.47.

8-14

Power Systems

Field data on the return-stroke velocity are limited. Lundholm (1957) and Rusck (1958) proposed the following empirical relationship between the return-stroke peak current and its velocity from the available field data: v=

c 1 + (500/I p )

(m/s)

(8.57)

where c is the velocity of electromagnetic fields in free space = 3 × 108 m/s Ip is the return-stroke peak current, kA The ground flash density, ng (number of flashes to ground per km2 per year), varies regionally and seasonally. However, the average ground flash density maps around the globe are available to estimate lightning-caused outages.

8.8 Estimation of Outage Rates Caused by Nearby Lightning Strokes The knowledge of the basic impulse insulation level (BIL) is essential for estimating the outage rate of an overhead power line. With this knowledge, the electrogeometric model is constructed to estimate the attractive area (Figure 8.7). According to the electrogeometric model, the striking distance of a lightning stroke is proportional to the return-stroke current. The following relation is used to estimate this striking distance, rs: rs = 8I p0.65 (m)

(8.58)

where Ip in kA is the peak of the return-stroke current. In the cross-sectional view of Figure 8.7, a horizontal line (representing a plane) is drawn at a distance of rs meters from the ground plane corresponding to the return-stroke current, Ip. A circular arc is drawn with its center on the conductor, P, and rs as radius. This represents a cylinder of attraction above the line conductor. The circular arc and the horizontal line intersect at points A and B. The strokes with I = Ip falling between A and B will strike the conductor, resulting in direct strokes; those falling outside AB will hit the ground, inducing voltages on the line. The horizontal projection of A or B is yo1, which is given by y o1 = rs for rs ≤ hp

y o1 = rs2 − (rs − hp )2

for rs > hp

A

rs P hp

(8.59a)

yo1

B

rs

FIGURE 8.7 Electrogeometric model for estimating the least distance of ground strike.

(8.59b)

8-15

Overvoltages Caused by Indirect Lightning Strokes

and yo1 is the shortest distance of a lightning stroke of the given return-stroke current from the overhead line, which will result in a flash to ground. To compute the outage rate, the return-stroke current, Ip, is varied from 1 to 200 kA in steps of 0.5 kA (Chowdhuri, 1989c, 2004). The current front time, tf, is varied from 0.5 to 10.5 μs in steps of 0.5 μs. At each current level, the shortest possible distance of the stroke, yo1, is computed from Equation 8.59. Starting at tf = 0.5 μs, the induced voltage is calculated as a function of time and compared with the given BIL of the line. If the BIL is not exceeded, then the next higher level of current is chosen. If the BIL is exceeded, then the lateral distance of the stroke from the line, y, is increased by Δy (e.g., 1 m), and the induced voltage is recalculated and compared with the BIL of the line. The lateral distance, y, is progressively increased until the induced voltage does not exceed BIL. This distance is called yo2. For the selected Ip and tf, the induced voltage will then exceed the BIL of the line and cause line flashover, if the lightning stroke hits the ground between yo1 and yo2 along the length of the line. For a 100 km sector of the line, the attractive area, A, will be (Figure 8.8)

A = 0.2( y o2 − y o1 ) km2

(8.60)

The joint probability density function, p(Ip, tf ), is then computed from Equation 8.56 for the selected Ip − tf combination. If ng is the ground flash density of the region, the expected number of flashovers per 100 km per year for that particular Ip − tf combination will be

nfo = p(I p , t f ) ⋅ ∆I p ⋅ ∆t f ⋅ ng ⋅ A

(8.61)

where ΔIp is the current step Δtf is the front-time step The front time, tf, is then increased by Δtf = 0.5 μs to the next step, and nfo for the same current but with the new tf is computed and added to the previous nfo. Once tf = 10.5 μs is reached, the return-stroke current is increased by ΔIp = 0.5 kA, and the whole procedure is repeated until the limits Ip = 200 kA and tf = 10.5 μs are reached. The cumulative nfo will then give the total number of expected line flashovers per 100 km per year for the selected BIL. The lightning-induced outage rates of the horizontally configured three-phase overhead line of Chapter 7 are plotted in Figure 8.9. The effectiveness of the shield wire, as shown in the figure, is optimistic, bearing in mind that the shield wire was assumed to be held at ground potential. The shield wire

A1 yo2 yo1 Overhead line yo1 yo2 A1

FIGURE 8.8 Attractive area of lightning ground flash to cause line flashover. A = 2A1 = 0.2(yo2 − yo1) km2.

8-16

Power Systems

No. of outages/100 km/year

10,000 1,000

No shield wire Two shield wires

100 10 1 0.1 0.01

50

100

150

200

250

300

350

400

450

500

550

600

Insulation withstand voltage (kV)

FIGURE 8.9 Outage rates of overhead line by indirect strokes vs. BIL.

No. of outages/100 km/year

10,000

Direct stroke Indirect stroke Total outage rate

1,000 100 10 1 0.1

50

100

150

(a)

200

250

300

350

400

450

500

550

600

550

600

Insulation withstand voltage (kV)

1,000

No. of outages/100 km/year

Direct stroke 100

Total outage rate 10 1 0.1 0.01

(b)

Indirect stroke

50

100

150

200

250

300

350

400

450

500

Insulation withstand voltage (kV)

FIGURE 8.10 Total outage rates of overhead line vs. BIL. (a) Unshielded line, and (b) shielded line.

8-17

Overvoltages Caused by Indirect Lightning Strokes

will not be held at ground potential under transient conditions. Therefore, the effectiveness of the shield wire will be less than the idealized case shown in Figure 8.9.

8.9 Estimation of Total Outage Rates Outages of an overhead power line may be caused by both direct and indirect lightning strokes, especially if the voltage withstand level of the line is low, for example, power distribution lines. Outages caused by direct strokes are caused by either backflash (i.e., lightning hitting the tower or the shield wire) or by shielding failure if lightning hits one of the phase conductors. Outages caused by direct strokes were discussed in Chapter 7. Outages caused by indirect strokes are caused by lightning strokes hitting the nearby ground, which is discussed in this chapter. Therefore, information from both these chapters will be required to estimate the total outage rates of overhead lines caused by lightning. The outage rates of the three-phase horizontally configured overhead line of Chapter 7 are shown in Figure 8.10a for the unshielded line and in Figure 8.10b for the shielded line. The effects of the variation of different parameters of the power line and also of the lightning stroke should also be considered in estimating the total outages of the line (Chowdhuri, 1989b). The power lines will be of finite length, terminated by transformers, underground cables, and lightning arresters. These terminations will generate multiple reflections. Therefore, power lines of finite lengths should also be considered for specific applications (Chowdhuri, 1991).

8.A Appendix A: Voltage Induced by Linearly Rising and Falling Return-Stroke Current V (x ,t ) = V1(x , t )u(t − t o ) + V2 (x , t )u(t − t of )

where

V1(x , t ) =

30α1hp 30α 2hp f12 f12a + 0.5n( f13a ) + 0.5n( f13 ) ; V2 (x , t ) = − bo ⋅ n bo ⋅ n βc f11 f11a βc bo = 1 − β2 ; t of = t o + t f ; t tf = t − t f

f1 = m1 + (ct − x )2 − y o2 ;

f 2 = m1 − (ct − x )2 + y o2

f 3 = m0 + (ct − x )2 + y o2 ;

f 4 = m0 + (ct o − x )2 − y o2

f 5 = n1 + (ct + x )2 − y o2 ;

f 6 = n1 − (ct + x )2 + y o2

f 7 = no − (ct o + x )2 + y o2 ;

f 8 = no + (ct o + x )2 − y o2

f 9 = bo (β2 x 2 + y o2 ) + β2c 2t 2 (t + β2 ); f11 =

c 2t 2 − x 2 ; y o2

f12 =

f10 = 2β2ct β2c 2t 2 + bo (x 2 + y o2 )

f 9 − f10 ; bo2 y o2

f13 =

f1 ⋅ f 3 ⋅ f 5 ⋅ f 7 f2 ⋅ f 4 ⋅ f6 ⋅ f8

8-18

Power Systems

f1a = m1a + (ct tf − x )2 − y o2 ;

f 3a = f 3 ;

f 5a = n1a + (ct tf + x )2 − y o2 ;

f 9a = bo (β2 x 2 + y o2 ) + β2c 2t tf2 (1 + β2 );

f11a =

f 4a = f 4 ;

c 2t tf2 − x 2 ; y o2

f12a =

f 2a = m1a − (ct tf − x )2 + y o2 f 7a = f 7 ;

f 8a = f 8

f 6a = n1a − (ct tf + x )2 + y o2 f10a = 2β2ct tf β2c 2t tf2 + bo (x 2 + y o2 )

f 9a − f10a ; bo2 y o2

f13a =

f1a ⋅ f 3a ⋅ f 5a ⋅ f 7a f 2a ⋅ f 4 a ⋅ f 6a ⋅ f 8a

m0 = [(ct o − x )2 + y o2 ]2 + 4hc2 (ct o − x )2 ; m1 = [(ct − x )2 + y o2 ]2 + 4hc2 (ct − x )2 n0 = [(ct o + x )2 + y o2 ]2 + 4hc2 (ct o + x )2 ; n1 = [(ct + x )2 + y o2 ]2 + 4hc2 (ct + x )2 m1a = [(ct tf − x )2 + y o2 ]2 + 4hc2 (ct tf − x )2 ; n1a = [(ct tf + x )2 + y o2 ]2 + 4hc2 (ct tf + x )2

References Agrawal, A.K., Price, H.J., and Gurbaxani, S.H., Transient response of multiconductor transmission lines excited by a nonuniform electromagnetic field, IEEE Trans. Electromagn. Compat., EMC-22, 119–129, 1980. Anderson, R.B. and Eriksson, A.J., Lightning parameters for engineering applications, Electra, 69, 65–102, 1980. Chowdhuri, P., Analysis of lightning-induced voltages on overhead lines, IEEE Trans. Power Delivery, 4, 479–492, 1989a. Chowdhuri, P., Parametric effects on the induced voltages on overhead lines by lightning strokes to nearby ground, IEEE Trans. Power Delivery, 4, 1185–1194, 1989b. Chowdhuri, P., Estimation of flashover rates of overhead power distribution lines by lightning strokes to nearby ground, IEEE Trans. Power Delivery, 4, 1982–1989, 1989c. Chowdhuri, P., Lightning-induced voltages on multiconductor overhead lines, IEEE Trans. Power Delivery, 5, 658–667, 1990. Chowdhuri, P., Response of overhead lines of finite length to nearby lightning strokes, IEEE Trans. Power Delivery, 6, 343–351, 1991. Chowdhuri, P., Electromagnetic Transients in Power Systems, 2nd edn., Research Studies Press, Baldock, Hertfordshire, U.K., 2004. Chowdhuri, P. and Gross, E.T.B., Voltage surges induced on overhead lines by lightning strokes, Proc. IEE (U.K.), 114, 1899–1907, 1967. Chowdhuri, P. and Gross, E.T.B., Voltages induced on overhead multiconductor lines by lightning strokes, Proc. IEE (U.K.), 116, 561–565, 1969. Chowdhuri, P., Li, S., and Yan, P., Review of research on lightning-induced voltages on an overhead line, IEE Proc.—Gener. Transm. Distrib.(U.K.), 148, 91–95, 2001.

Overvoltages Caused by Indirect Lightning Strokes

8-19

Cinieri, E. and Fumi, A., The effect of the presence of multiconductors and ground wires on the atmospheric high voltages induced on electrical lines (in Italian), L’Energia Elettrica, 56, 595–601, 1979. Eriksson, A.J., Notes on lightning parameters for system performance estimation, CIGRE Note 33–86 (WG33-01) IWD, 15 July 1986. Fortescue, C.L., Direct strokes—Not induced surges—Chief cause of high-voltage line flashover, The Electric Journal, 27, 459–462, 1930. Haldar, M.K. and Liew, A.C., Alternative solution for the Chowdhuri-Gross model of lightning-induced voltages on power lines, Proc. IEE (U.K.), 135, 324–329, 1988. IEEE = PES Task Force 15.09, Parameters of lightning strokes: A review, IEEE Trans. Power Delivery, 20, 346–358, 2005. Lundholm, R., Induced overvoltage-surges on transmission lines and their bearing on the lightning performance at medium voltage networks, Trans. Chalmers Univ. Technol., Gothenburg, Sweden, 188, 1–117, 1957. Morse, P.M. and Feshbach, H., Methods of Theoretical Physics, Vol. 1, McGraw-Hill, New York, 1950, Chap. 7. Rusck, S., Induced lightning over-voltages on power-transmission lines with special reference to the overvoltage protection of low-voltage networks, Trans. R. Inst. Tech., Stockholm, Sweden, 120, 1–118, 1958.

9 Switching Surges Stephen R. Lambert Shawnee Power Consulting, LLC

9.1 Transmission Line Switching Operations...................................... 9-1 9.2 Series Capacitor Bank Applications................................................ 9-5 9.3 Shunt Capacitor Bank Applications................................................ 9-5 9.4 Shunt Reactor Applications.............................................................. 9-5 References.......................................................................................................9-6

Switching surges occur on power systems as a result of instantaneous changes in the electrical configuration of the system, and such changes are mainly associated with switching operations and fault events. These overvoltages generally have crest magnitudes which range from about 1 to 3 pu for phaseto-ground surges and from about 2 to 4 pu for phase-to-phase surges (in pu on the phase to ground crest voltage base) with higher values sometimes encountered as a result of a system resonant condition. Waveshapes vary considerably with rise times ranging from 50 μs to thousands of μs and times to halfvalue in the range of hundreds of μs to thousands of μs. For insulation testing purposes, a waveshape having a time to crest of 250 μs with a time to half-value of 2000 μs is often used. The following addresses the overvoltages associated with switching various power system devices. Possible switching surge magnitudes are indicated, and operations and areas of interest that might warrant investigation when applying such equipment are discussed.

9.1 Transmission Line Switching Operations Surges associated with switching transmission lines (overhead, SF6, or cable) include those that are generated by line energizing, reclosing (three phase and single phase operations), fault initiation, line dropping (deenergizing), fault clearing, etc. During an energizing operation, for example, closing a circuit breaker at the instant of crest system voltage results in a 1 pu surge traveling down the transmission line and being reflected at the remote, open terminal. The reflection interacts with the incoming wave on the phase under consideration as well as with the traveling waves on adjacent phases. At the same time, the waves are being attenuated and modified by losses. Consequently, it is difficult to accurately predict the resultant waveshapes without employing sophisticated simulation tools such as a transient network analyzer (TNA) or digital programs such as the Electromagnetic Transients Program (EMTP). Transmission line overvoltages can also be influenced by the presence of other equipment connected to the transmission line—shunt reactors, series or shunt capacitors, static var systems, surge arresters, etc. These devices interact with the traveling waves on the line in ways that can either reduce or increase the severity of the overvoltages being generated. When considering transmission line switching operations, it can be important to distinguish between “energizing” and “reclosing” operations, and the distinction is made on the basis of whether the line’s inherent capacitance retains a trapped charge at the time of line closing (reclosing operation) or whether

9-1

9-2

Power Systems

no trapped charge exists (an energizing operation). The distinction is important as the magnitude of the switching surge overvoltage can be considerably higher when a trapped charge is present; with higher magnitudes, insulation is exposed to increased stress, and devices such as surge arresters will, by necessity, absorb more energy when limiting the higher magnitudes. Two forms of trapped charges can exist—DC and oscillating. A trapped charge on a line with no other equipment attached to the line exists as a DC trapped charge, and the charge can persist for some minutes before dissipating (Beehler, 1964). However, if a transformer (power or wound potential transformer) is connected to the line, the charge will decay rapidly (usually in less than 0.5 s) by discharging through the saturating branch of the transformer (Marks, 1969). If a shunt reactor is connected to the line, the trapped charge takes on an oscillatory waveshape due to the interaction between the line capacitance and the reactor inductance. This form of trapped charge decays relatively rapidly depending on the Q of the reactor, with the charge being reduced by as much as 50% within 0.5 s. Figures 9.1 and 9.2 show the switching surges associated with reclosing a transmission line. In Figure 9.1 note the DC trapped charge (approximately 1.0 pu) that exists prior to the reclosing operation (at 20 μs). Figure 9.2 shows the same case with an oscillating trapped charge (a shunt reactor was present on the line) prior to reclosing. Maximum surges were 3.0 for the DC trapped charge case and 2.75 pu for the oscillating trapped charge case (both occurred on phase c). The power system configuration behind the switch or circuit breaker used to energize or reclose the transmission line also affects the overvoltage characteristics (shape and magnitude) as the traveling wave interactions occurring at the junction of the transmission line and the system (i.e., at the circuit breaker) as well as reflections and interactions with equipment out in the system are important. In general, a stronger system (higher short circuit level) results in somewhat lower surge magnitudes than a weaker system, although there are exceptions. Consequently, when performing simulations to predict overvoltages, it is usually important to examine a variety of system configurations (e.g., a line out of service or contingencies) that might be possible and credible.

(10 µs/div.)

FIGURE 9.1 DC trapped charge.

9-3

Switching Surges

(10 µs/div.)

FIGURE 9.2 Oscillating trapped charge.

Single phase switching as well as three phase switching operations may also need to be considered. On EHV transmission lines, for example, most faults (approximately 90%) are single phase in nature, and opening and reclosing only the faulted phase rather than all three phases, reduces system stresses. Typically, the overvoltages associated with single phase switching have a lower magnitude than those that occur with three phase switching (Koschik et al., 1978). Switching surge overvoltages produced by line switching are statistical in nature—that is, due to the way that circuit breaker poles randomly close (excluding specially modified switchgear designed to close on or near voltage zero), the instant of electrical closing may occur at the crest of the system voltage, at voltage zero, or somewhere in between. Consequently, the magnitude of the switching surge varies with each switching event. For a given system configuration and switching operation, the surge voltage magnitude at the open end of the transmission line might be 1.2 pu for one closing event and 2.8 pu for the next (Hedman et al., 1964; Johnson et al., 1964), and this statistical variation can have a significantly impact on insulation design (see Chapter 14 on insulation coordination). Typical switching surge overvoltage statistical distributions (160 km line, 100 random closings) are shown in Figures 9.3 and 9.4 for phase-to-ground and phase-to-phase voltages (Lambert, 1988), and the surge magnitudes indicated are for the highest that occurred on any phase during each closing. With no surge limiting action (by arresters or circuit breaker preinsertion resistors), phase-to-ground surges varied from 1.7 to 2.15 pu with phase-to-phase surges ranging from 2.2 to 3.7 pu. Phase-to-phase surges can be important to line-connected transformers and reactors as well as to transmission line phase-to-phase conductor separation distances when line uprating or compact line designs are being considered. Figure 9.3 also demonstrates the effect of the application of surge arresters on phase-to-ground surges, and shows the application of resistors preinserted in the closing sequence of the circuit breaker (400 Ω for 5.56 ms) is even more effective than arresters in reducing surge magnitude. The results shown on

9-4

Power Systems

Surge magnitude (pu)

3.0

No arrester or resistor

2.0

Arresters Resistors

1.0 0.01

1

10 50 90 99 Percent exceeding ordinate value

99.99

FIGURE 9.3 Phase-to-ground overvoltage distribution.

Surge magnitude (pu) (φ – φ pu on φ – G base)

4.0

No arrester or resistor 3.0

Arresters

Resistors

2.0 0.01

1

10

50

90

99

99.99

FIGURE 9.4 Phase-to-phase overvoltage distribution.

Figure 9.4, however, indicate that while resistors are effective in limiting phase-to-phase surges, arresters applied line to ground are generally not very effective at limiting phase-to-phase overvoltages. Line dropping (deenergizing) and fault clearing operations also generate surges on the system, although these typically result in phase-to-ground overvoltages having a maximum value of 2–2.2 pu. Usually the concern with these operations is not with the phase-to-ground or phase-to-phase system voltages, but rather with the recovery voltage experienced by the switching device. The recovery voltage is the voltage which appears across the interrupting contacts of the switching device (a circuit breaker for example) following current extinction, and if this voltage has too high a magnitude, or in some instances rises to its maximum too quickly, the switching device may not be capable of successfully interrupting. The occurrence of a fault on a transmission line also can result in switching surge type overvoltages, especially on parallel lines. These voltages usually have magnitudes on the order of 1.8–2.2 pu and are usually not a problem (Kimbark and Legate, 1968; Madzarevic et al., 1977).

9-5

Switching Surges

Bank 1

Bank 2

FIGURE 9.5 Voltage magnification circuit.

9.2 Series Capacitor Bank Applications Installation of a series capacitor bank in a transmission line (standard or thyristor controlled) has the potential for increasing the magnitude of phase-to-ground and phase-to-phase switching surge overvoltages due to the trapped charges that can be present on the bank at the instant of line reclosing. In general, surge arresters limit the phase-to-ground and phase-to-phase overvoltages to acceptable levels; however, one problem that can be serious is the recovery voltage experienced by circuit breakers when clearing faults on a series compensated line. Depending the bank’s characteristics and on fault location with respect to the bank’s location, a charge can be trapped on the bank, and this trapped charge can add to the surges already being generated during the fault clearing operation (Wilson, 1972). The first circuit breaker to clear is sometimes exposed to excessive recovery voltages under such conditions.

9.3 Shunt Capacitor Bank Applications Energizing a shunt capacitor bank typically results in maximum overvoltages of about 2 pu or less. However, there are two conditions where significant overvoltages can be generated. One involves a configuration (shown on Figure 9.5) where two banks are separated by a significant inductance (e.g., a transformer) (Schultz et al., 1959). When one bank is switched, if the system inductance and bank 1 capacitance has the same natural frequency as that of the transformer leakage inductance and the bank 2 capacitance, then a voltage magnification can take place. Another configuration that can result in damaging overvoltages involves energizing a capacitor bank with a transformer terminated transmission line radially fed from the substation at which the capacitor bank is located (Jones and Fortson, 1985). During bank switching, phase-to-phase surges are imposed on the transformer, and because these are not very well suppressed by the usual phase-toground application of surge arresters, transformer failures have been known to result. Various methods to reduce the surge magnitude have included the application of controlled circuit breaker closing techniques (closing near voltage zero), and resistors or reactors preinserted in the closing sequence of the switching devices. Restriking of the switching device during bank deenergizing can result in severe line-to-ground overvoltages of 3–5 pu or more (rarely) (Johnson et al., 1955; Greenwood, 1971). Surge arresters are used to limit the voltages to acceptable levels, but at higher system voltages, the energy discharged from the bank into the arrester can exceed the arrester’s capability.

9.4 Shunt Reactor Applications Switching of shunt reactors (and other devices characterized as having small inductive currents such as transformer magnetizing currents, motor starting currents, etc.) can generate high phase-to-ground overvoltages as well as severe recovery voltages (Greenwood, 1971), especially on lower voltage equipment such as reactors applied on the tertiary of transformers. Energizing the devices seldom generates high overvoltages, but overvoltages generated during deenergizing, as a result of current chopping by

9-6

Power Systems

the switching device when interrupting the small inductive currents, can be significant. Neglecting damping, the phase-to-ground overvoltage magnitude can be estimated by

V =i

L C

(9.1)

where i is the magnitude of the chopped current (0 to perhaps as high as 10 A or more) L is the reactor’s inductance C is the capacitance of the reactor (on the order of a few thousand picofarads) When C is small, especially likely with dry-type reactors often used on transformer tertiaries, the surge impedance term can be large, and hence the overvoltage can be excessive. To mitigate the overvoltages, surge arresters are sometimes useful, but the application of a capacitor on the terminals of the reactor (or other equipment) have a capacitance on the order of 0.25–0.5 μF is very helpful. In the equation above, note that if C is increased from pF to μF, the surge impedance term is dramatically reduced, and hence the voltage is reduced.

References Beehler, J.E., Weather, corona, and the decay of trapped energy on transmission lines, IEEE Trans. Power Appar. Syst. 83, 512, 1964. Greenwood, A., Electrical Transients in Power Systems, John Wiley & Sons, New York, 1971. Hedman, D.E., Johnson, I.B., Titus, C.H., and Wilson, D.D., Switching of extra-high-voltage circuits, II—Surge reduction with circuit breaker resistors, IEEE Trans. Power Appar. Syst. 83, 1196, 1964. Johnson, I.B., Phillips, V.E., and Simmons, Jr., H.O., Switching of extra-high-voltage circuits, I—system requirements for circuit breakers, IEEE Trans. Power Appar. Syst. 83, 1187, 1964. Johnson, I.B., Schultz, A.J., Schultz, N.R., and Shores, R.R., Some fundamentals on capacitance switching, AIEE Trans. Power Appar. Syst. PAS-74, 727, 1955. Jones, R.A. and Fortson, Jr., H.S., Considerations of phase-to-phase surges in the application of capacitor banks, IEEE PES Summer Meeting, Vancouver, Canada, 1985, 85 SM 400–7. Kimbark, E.W. and Legate, A.C., Fault surge versus switching surge: A study of transient overvoltages caused by line-to-ground faults, IEEE Trans. Power Appar. Syst. PAS-87, 1762, 1968. Koschik, V., Lambert, S.R., Rocamora, R.G., Wood, C.E., and Worner, G., Long line single-phase switching transients and their effect on station equipment, IEEE Trans. Power Appar. Syst. PAS-97, 857, 1978. Lambert, S.R., Effectiveness of zinc oxide surge arresters on substation equipment probabilities of flashover, IEEE Trans. Power Delivery 3(4), 1928, 1988. Madzarevic, V., Tseng, F.K., Woo, D.H., Niebuhr, W.D., and Rocamora, R.G., Overvoltages on EHV transmission lines due to fault and subsequent bypassing of series capacitors, IEEE PES Winter Meeting, New York, January 1977, F77 237–1. Marks, L.W., Line discharge by potential transformers, IEEE Trans. Power Appar. Syst. PAS-88, 293, 1969. Schultz, A.J., Johnson, J.B., and Schultz, N.R., Magnification of switching surges, AIEE Trans. Power Appar. Syst. 77, 1418, 1959. Wilson, D.D., Series compensated lines—voltages across circuit breakers and terminals caused by switching, IEEE PES Summer Meeting, San Francisco, CA, 1972, T72 565–0.

10 Very Fast Transients 10.1 Origin of VFT in GIS...................................................................... 10-2 10.2 Propagation of VFT in GIS............................................................ 10-3 Internal Transients • External Transients

10.3 Modeling Guidelines and Simulation.......................................... 10-7 Computation of Internal Transients • Computation of TEV • Testing and Simulation • Statistical Calculation • Validation

Juan A. Martinez-Velasco Universitat Politecnica de Catalunya

10.4 Effects of VFT on Equipment...................................................... 10-16 SF6 Insulation • Transformers • Disconnectors and Breakers • Enclosure and Cable Interfaces • Bushings • Secondary Equipment

References................................................................................................... 10-17

Electromagnetic transient phenomena in power systems are caused by switching operations, faults, and other disturbances, such as lightning strokes. They may appear with a wide range of frequencies that vary from DC to several MHz and occur on a time scale that goes from microseconds to several cycles. Due to frequency-dependent behavior of power components and difficulties for developing models accurate enough for a wide frequency range, the frequency ranges are classified into groups, with overlapping between them. An accurate mathematical representation of each power component can generally be developed for a specific frequency range (CIGRE, 1990; IEC, 2006). Very fast transients (VFT), also known as very fast front transients, belong to the highest frequency range of transients in power systems. According to the classification proposed by the CIGRE Working Group 33-02, VFT may vary from 100 kHz up to 50 MHz (1990). According to IEC 60071-1, the shape of a very fast front overvoltage is “usually unidirectional with time to peak 0 2π 2 AWire

(11.12)

Equation 11.12 is valid for most electrode shapes, such as combinations of rings, grids, and rods. When several (N) radial wires meet at a point, the contact resistance coefficient K in Equation 11.10 should use K=

1 N 2 ATotal ln , K >0 2π 8 AWire

(11.13)

where N is the number of wires meeting at a point ATotal is the surface area of the pattern made by the radial array AWire is the surface area of the wires themselves

11.9 Numerical Treatment of Complex Electrode Shapes Solving for the combined resistance of a number of individual electrode elements, such as four foundations of a transmission tower along with some horizontal wires and ground rods, can take several approaches. The geometric resistance of the entire electrode can be computed, and the contact resistance correction for a wire frame approximation can be added as in Equation 11.11. However, in some cases such as solid cylinders of concrete, the “length” of wire to use is not obvious, and also, the value of K varies depending on whether the electrode is tightly meshed or open. A second possibility is to develop a matrix of self- and mutual-resistance values of each solid, conducting element. The geometric resistances defined in Equation 11.10 give the self-resistance Rii. The mutual resistance between the centroids of two objects is calculated from their separation dij:

Rij =

ρ 2πdij

(11.14)

A symmetrical matrix of self- and mutual resistances is multiplied by a vector of current values to obtain the potentials at each electrode. The matrix can be built, inverted, and solved in Excel as follows.

11-11

Transmission System Transients: Grounding 3 6 |- 5 m -| 9 1 4

2 7

8 5 R = 2.16 Ω

3 |- 5 m -| 1

2

4

5

R = 3.01 Ω

2 |- 5 m -| 1 4

2 |- 5 m -| 1

R = 4.52 Ω

1

R = 6.43 Ω

R = 11.85 Ω

FIGURE 11.3 Numerical solutions of resistance for 10 m rods at 5 m grid spacing (IEEE Guide 80/2000).

In our case, we will consider five 3 m long rods of 5 mm radius in ρ = 320 Ωm soil, on a 5 m × 5 m grid as shown in Figure 11.3. The self-resistance is calculated from g = 10 m, A = 2πrl = 0.4 m2 in Equation 11.10 as 119 Ω. The distance between the center rod and all others is d1j = 7.07 m, d23 = d34 = d45 = d25 = 10 m, and d24 = d35 = 14.1 m. The mutual resistances are given by Rij = ρ/(2πdij) from Equation 11.14 as follows: Resistances Rij In Ω (ρ = 320 Ωm)

Distances Dij in m 0.01 7.07 7.07 7.07 7.07

7.07 0.01 10.00 14.14 10.00

7.07 10.00 0.01 10.00 14.14

7.07 14.14 10.00 0.01 10.00

7.07 10.00 14.14 10.00 0.01

119.30 7.20 7.20 7.20 7.20

7.20 119.30 5.09 3.60 5.09

7.20 5.09 119.30 5.09 3.60

7.20 3.60 5.09 119.30 5.09

7.20 5.09 3.60 5.09 119.30

In Excel, a square matrix can be inverted by highlighting a blank space of the same size (5 × 5 cells in this case), typing in the formula = MINVERSE(A1:E5) where the original Rij matrix resides in cells A1:E5, and pressing the F2 key and then at the same time. The inverted matrix can be multiplied by a vector of unit potentials (1s) to obtain the currents in each electrode. The unit potential divided by the sum of these currents gives the resistance of the combined electrode, including self and mutual effects. The following table completes the numerical example. Inverse of Rij Matrix (×1000) 8.49 −0.46 −0.46 −0.46 −0.46

−0.46 8.44 −0.31 −0.20 −0.31

−0.46 −0.31 8.44 −0.31 −0.20

−0.46 −0.20 −0.31 8.44 −0.31

−0.46 −0.31 −0.20 −0.31 8.44

Potential (V)

Current = MMULT([Rij],VI)

1 1 1 1 1

6.65 mA 7.15 mA 7.15 mA 7.15 mA 7.15 mA 35.27 mA 1/.03527 = 28.4 Ω

Sum

To multiply the inverted resistance matrix (say in locations A10:E15) by a vector of unit potentials (say in locations G10:G15), the cells I10:I15 would be highlighted, the formula = MMULT(A10:E15,G10:G15) would be entered, and the F2 key would be pressed and then pressed at the same time. The currents are summed to a total of 35.27 mA, giving a resistance of 28.4 Ω. The five resistances in parallel, ignoring mutual effects, would have a resistance of 23.9 Ω, so the effect at a separation of 5 m is still significant. A series of numerical calculations are found for a similar set of electrodes in (IEEE Guide 80, 2000, p. 183) for 10 m rods on 5 m grid spacing in 100 Ωm soil. The relative reduction in resistance is largest when the second electrode is added, and additional nearby rods are seen to be less effective.

11.10 Treatment of Multilayer Soil Effects Generally, the treatment of footing resistance in lightning calculations considers a homogeneous soil with a finite conductivity. This treatment, however, seldom matches field observations, particularly in areas where grounding is difficult. Under these conditions, a thin “overburden” layer of conducting clay,

11-12

Power Systems

Ratio: effective to upper layer resistivity, ρe/ρ1

100

ρ2/ρ1 = 100 ρ2/ρ1 = 50 ρ2/ρ1 = 20

10

ρ2/ρ1 = 10 ρ2/ρ1 = 5 ρ2/ρ1 = 2

1

ρ2/ρ1 = 1 ρ2/ρ1 = 0.5 ρ2/ρ1 = 0.2 ρ2/ρ1 = 0.1

0.1

0.01 0.01

ρ2/ρ1 = 0.05 ρ2/ρ1 = 0.02 ρ2/ρ1 = 0.01

0.1

1

10

100

1000

Ratio: Ring electrode radius/upper layer depth

FIGURE 11.4 Relative effective resistivity versus ratio of electrode radius to upper-layer depth, with resistivity ratio as a parameter.

till, or gravel often rests on top of insulating rock. The distribution of resistivity values for a particular overburden material and condition can be narrow, with standard deviations usually less than 10%. However, the variation of overburden depth with distance can be large. Airborne electromagnetic survey techniques at multiple frequencies in the 10–100 kHz range offer a cost-effective method of reconnaissance of the overburden parameters of resistivity and depth. Once a resistivity survey has established an upper-layer resistivity ρ1, a layer depth d, and a lowerlayer resistivity ρ2, the equivalent resistivity ρe can be computed. For a disk-like electrode buried just below the surface, ρe from the elliptic-integral solutions of (Zaborsky, 1955) can approximated with better than 5% accuracy by the following empirical equations: ρe = ρ1

1 + C(ρ2 /ρ1 )(r /d ) 1 + C(r /d )

1 ρ1 ≥ ρ2 , 1.4 + (ρ /ρ )0.8 2 1 C= 1 ρ1 < ρ2 , 1.4 + (ρ2 /ρ1 )0.8 + ((ρ2 /ρ1 )(r /d ))0.5

(11.15)

The ratio of effective resistivity to upper-layer resistivity varies with the ratio of electrode radius s to upper-layer depth d, with a small ratio of s/d giving a ratio of unity as shown in Figure 11.4. Normally, electrode penetration through an upper layer would only be desirable in extreme examples of Case 1 (ρ1 >> ρ2). Rather than recomputing the effective resistivity with revised image locations, the effects of the upper layer can be neglected, with the connection through ρ1 providing only series self-inductance.

11.11 Layer of Finite Thickness over Insulator A simpler two-layer soil treatment is appropriate for Case 2 when ρ2 >> ρ1, or equivalently the reflection coefficient from upper to lower layer Γ12 approaches 1. Under these conditions, the following summation in Equation 11.16 describes the resistance of a single hemisphere of radius s in a finitely conducting slab with resistivity ρ1 and thickness h:

11-13

Transmission System Transients: Grounding

Resistance Ratio: Finite to infinite depth

6

Keller, r = 10 m, 10,000 terms

Loyka estimate, 105 d

5

Keller, r = 3 m, 10,000 terms

5 Loyka estimate, 10 d

Keller, r = 1 m, 10,000 terms

Loyka estimate, 105 d

4

Keller, r = 0.3 m, 10,000 terms

Loyka estimate, 105 d

3 2 1 0

1

10

100

1000

Depth of upper layer, m

FIGURE 11.5 Asymptotic behavior of hemispherical electrode in conducting layer over insulator.

Γ12 =

ρ2 − ρ1 ρ ≈ 1 for ρ2 >> ρ1 RKeller = 1 1 + 2 ⋅ 2πs ρ2 + ρ1

RLoyka ≈

∞

∑ n =1

1 + (2nh/s)2 n Γ12

50, 000 ρ1 s 1 + 1 + (105 h/2h)2 ρ1 s ≈ 1 + ln 1 + ln 2πs h 1 + 1 + (s /2h)2 2πs h 1 + 1 + (s /2h)2

(11.16)

For a reflection coefficient Γ12 = 1, the Keller series in Equation 11.16 becomes the harmonic series and converges slowly. Loyka (1999) offers a convenient approximation to the sum, developed using the potential from a second hemisphere, located sufficiently far away (e.g., 105d, where d is the layer thickness) to have no influence. Figure 11.5 shows that the Loyka approximation to the Keller series is good for a wide range of transmission line grounding applications, and that finite upper layer depth has a strong influence on the resistance of the electrodes.

11.12 Treatment of Soil Ionization Under high electric fields, air will ionize and become effectively a conductor. The transient electric fields needed to ionize air in small volumes of soil, or to flashover across the soil surface, are typically between 100 and 1000 kV/m (Korsuncev, 1958; Liew and Darveniza, 1974; Oettle, 1988). Considering that the potential rise on a small ground electrode can reach 1 MV, the origins of 10 m furrows around small (inadequate) ground electrodes after lightning strikes become clear. Surface arcing activity is unpredictable and may transfer lightning surge currents to unprotected facilities. Thus, power system ground electrodes for lightning protection should have sufficient area and multiplicity to limit ionization. Korsuncev used similarity analysis to relate dimensionless ratios of s, ρ, resistance R, current I, and critical breakdown gradient Eo as follows:

Π1 =

Rs ρ

with Π1O =

1 2πes 2 ln 2π A

(11.17)

11-14

Power Systems

Π2 =

ρI Eo s 2

(

(11.18)

− 0.31

Π1 = MIN Π1O , 0.26 ⋅ Π 2

R=

)

ρΠ1 s

(11.19)

(11.20)

where s is the three-dimensional distance from the center to the furthest point on the electrode, m I is the electrode current, A ρ is the resistivity, Ωm Eo is the critical breakdown gradient of the soil, usually 300–1000 kV/m R is the resistance of the ground electrode under ionized conditions The calculation of ionized electrode resistance proceeds as follows: • A value of Π1O is calculated from Equation 11.17. This un-ionized value will range from Π1O = 0.159 (for a hemisphere) to 1.07 for a 3 m long, 0.01 m radius cylinder. • A value of Π2 is calculated from Equation 11.18. For a 3 m rod at 100 kA in 100 Ωm soil, with Eo of 300 kV/m, the value of Π2 = 3.7 is obtained. −0.31 • A value of Π1 is computed from Π 2 using 0.26 Π 2 . This value, from Equation 11.19, represents the fully ionized sphere with the gradient of Eo at the injected current; for the rod example, Π1 = 0.173. • If the ionized value of Π1 is not greater than Π1O, then there is not enough current to ionize the footing, so the unionized resistance from Π1O will be seen for the calculation of resistance in Equation 11.20. For the 3 m rod, ionization reduces the 36 Ω low-current resistance to 6 Ω at 100 kA. In two-layer soils with sparse electrodes, ionization effects will tend to reduce the contact resistance (from Equation 11.10) without altering the surface area A or characteristic dimension s. This will tend to reduce the influence of ionization, since the geometric resistance of the electrode is unchanged.

11.13 Design Process The inputs to the design process are the prospective lightning surge current, the local soil resistivity, and the dimensions of the tower base or trial electrode. The low-frequency geometric resistance of the electrode is computed from Equation 11.9 and double-checked with the estimate from Equations 11.15 and 11.20. If the electrode is compact (e.g., a driven rod), then the ionization effects should be estimated using Equations 11.18 and 11.19, with the reduced resistance under high-current conditions being given by Equation 11.20. If the electrode is distributed (e.g., a buried wire), then the ionized resistance will only reduce the contact resistance term in Equation 11.10, and for practical purposes it will be sufficient to disregard this term and consider only the geometric resistance estimated from Equation 11.8 or Equation 11.9. In each case, an apparent transient impedance of the ground electrode should be modeled as an equivalent inductance in series with the resistive rise associated with the geometric resistance. The fraction of surge current absorbed by a given electrode will change as its size is adjusted, making the design process iterative.

11-15

Transmission System Transients: Grounding

11.14 Design Recommendations It is prudent where possible to select an electrode size that does not rely on ionization for adequate transient performance. This can be achieved by inverting the design process as follows. • Establish Π1O, the shape coefficient of the electrode (0.16 for a hemisphere, 0.26 for a cube, and 0.27 for a disk). • Establish the value of ΠO2 that will just cause ionization (Chisholm and Janischewskyj 1989), inverting Equation 11.19 to obtain ΠO2 = 0.0131( Π1O )

−3.24

(11.21)

For the example of a disk electrode, the value of ΠO2 = 0.92 is obtained. • The extent (in this case radius) of the electrode s needed to prevent ionization in soil of resistivity ρ in Ωm with current I in kA and voltage gradient Eo in kV/m is given by s=

ρI Eo ΠO2

(11.22)

For ρ = 300 Ωm, I = 30 kA, and Eo of 400 kV/m, a disk or ring radius of s = 9.1 m will be sufficient to prevent ionization. The geometric resistance of this electrode from Equation 11.20 is 8.9 Ω. The following advice is especially relevant for transmission towers or other tall structures, where a low-impedance ground is needed to limit lightning-transient overvoltages: • Choose a wide, flat electrode shape rather than a long, thin shape. Four radial counterpoise wires of 60 m will be two to eight times more effective than a single counterpoise of 240 m under lightning surge conditions. • Take advantage of natural elements in the structure grounding, such as foundations and guy anchors, by planning for electrical connections and by extending radial wires outward from these points. • In rocky areas, use modern airborne techniques to survey resistivity and layer depth using several frequencies up to 100 kHz. Place towers where conductive covering material is deep. • Provide grounding staff with the tools and techniques to preestimate the amount of wire required for target footing impedance values, using simple interpretation of two-layer soil data. • Near areas where transferred lightning potentials could be dangerous to adjacent objects or systems, use sufficient electrode dimensions to limit ionization, that is to remain on the Π1O characteristic.

11.A Appendix A: Relevant IEEE and IEC Standards in Lightning and Grounding 11.A.1 A NSI/IEEE Std 80-1986: IEEE Guide for Safety in AC Substation Grounding Presents essential guidelines for assuring safety through proper grounding at ac substations at all voltage levels. Provides design criteria to establish safe limits for potential differences within a station, under fault conditions, between possible points of contact. Uses a step-by-step format to describe test methods, design and testing of grounding systems. Provides English translations of fundamental papers on grounding by such as Rüdenberg and Laurent that are not widely available.

11-16

Power Systems

11.A.2 A NSI/IEEE Std 80-2000: IEEE Guide for Safety in AC Substation Grounding Provides an improved methodology for interpreting two-layer soil resistivity and using the values in the design of ac substations. Provides methods for determining the maximum grid current at substations, some of which also predict the maximum fault currents available on lines close by. Provides a number of new worked examples in appendices.

11.A.3 IEC 62305: Protection against Lightning, Edition 2.0, 2010–12 The European Technical Committee 81 of the International Electrotechnical Commission (IEC, www.iec.ch) prepared an authoritative and comprehensive lightning protection standard in five parts as follows: • • • • •

Part 1, General principles Part 2, Risk management Part 3, Physical damage to structures and life hazard Part 4, Electrical and electronic systems within structures Part 5, Services (Edition 1)

The rationale in these reference documents forms a sound basis for national lightning protection standards of structures and wind turbines.

11.A.4 I EEE Std 81-1983: IEEE Guide for Measuring Earth Resistivity, Ground Impedance, and Earth Surface Potentials of a Ground System The present state of the technique of measuring ground resistance and impedance, earth resistivity, and potential gradients from currents in the earth, and the prediction of the magnitude of ground resistance and potential gradients from scale-model tests are described and discussed. Factors influencing the choice of instruments and the techniques for various types of measurements are covered. These include the purpose of the measurement, the accuracy required, the type of instruments available, possible sources of error, and the nature of the ground or grounding system under test. The intent is to assist the engineer or technician in obtaining and interpreting accurate, reliable data. The test procedures described promote the safety of personnel and property and prevent interference with the operation of neighboring facilities. The standard is under revision as of September 2010.

11.A.5 I EEE Std 81.2-1991: IEEE Guide for Measurement of Impedance and Safety Characteristics of Large, Extended, or Interconnected Grounding Systems Practical instrumentation methods are presented for measuring the ac characteristics of large, extended, or interconnected grounding systems. Measurements of impedance to remote earth, step and touch potentials, and current distributions are covered for grounding systems ranging in complexity from small grids (less than 900 m2) with only a few connected overhead or direct-burial bare concentric neutrals, to large grids (greater than 20,000 m2) with many connected neutrals, overhead ground wires (sky wires), counterpoises, grid tie conductors, cable shields, and metallic pipes. This standard addresses measurement safety; earth-return mutual errors; low-current measurements; power-system staged

Transmission System Transients: Grounding

11-17

faults; communication and control cable transfer impedance; current distribution (current splits) in the grounding system; step, touch, mesh, and profile measurements; the foot-equivalent electrode earth resistance; and instrumentation characteristics and limitations.

11.A.6 I EEE Std 367-1996: IEEE Recommended Practice for Determining the Electric Power Station Ground Potential Rise and Induced Voltage from a Power Fault Information for the determination of the appropriate values of fault-produced power station ground potential rise (GPR) and induction for use in the design of protection systems is provided. Included are the determination of the appropriate value of fault current to be used in the GPR calculation; taking into account the waveform, probability, and duration of the fault current; the determination of inducing currents, the mutual impedance between power and telephone facilities, and shield factors; the vectorial summation of GPR and induction; considerations regarding the power station GPR zone of influence; and communications channel time requirements for noninterruptible services. Guidance for the calculation of power station GPR and longitudinal induction (LI) voltages is provided, as well as guidance for their appropriate reduction from worst-case values, for use in metallic telecommunication protection design.

11.A.7 IEEE Std 524a-1993: IEEE Guide to Grounding during the Installation of Overhead Transmission Line Conductors—Supplement to IEEE Guide to the Installation of Overhead Transmission Line Conductors General recommendations for the selection of methods and equipment found to be effective and practical for grounding during the stringing of overhead transmission line conductors and overhead ground wires are provided. The guide is directed to transmission voltages only. The aim is to present in one document sufficient details of present day grounding practices and equipment used in effective grounding and to provide electrical theory and considerations necessary to safeguard personnel during the stringing operations of transmission lines.

11.A.8 I EEE Std 837-2002: IEEE Standard for Qualifying Permanent Connections Used in Substation Grounding Directions and methods for qualifying permanent connections used for substation grounding are provided. Particular attention is given to the connectors used within the grid system, connectors used to join ground leads to the grid system, and connectors used to join the ground leads to equipment and structures. The purpose is to give assurance to the user that connectors meeting the requirements of this standard will perform in a satisfactory manner over the lifetime of the installation provided, that the proper connectors are selected for the application, and that they are installed correctly. Parameters for testing grounding connections on aluminum, copper, steel, copper-clad steel, galvanized steel, stainless steel, and stainless-clad steel are addressed. Performance criteria are established, test procedures are provided, and mechanical, current–temperature cycling, freeze-thaw, corrosion, and fault-current tests are specified.

11.A.9 I EEE Std 1048-2003: IEEE Guide for Protective Grounding of Power Lines Guidelines are provided for safe protective grounding methods for persons engaged in deenergized overhead transmission and distribution line maintenance. They comprise state-of-the-art information

11-18

Power Systems

on protective grounding as currently practiced by power utilities in North America. The principles of protective grounding are discussed. Grounding practices and equipment, power-line construction, and ground electrodes are covered.

11.A.10 I EEE Std 1050-2004: IEEE Guide for Instrumentation and Control Equipment Grounding in Generating Stations Information about grounding methods for generating station instrumentation and control (I & C) equipment is provided. The identification of I & C equipment grounding methods to achieve both a suitable level of protection for personnel and equipment is included, as well as suitable noise immunity for signal ground references in generating stations. Both ideal theoretical methods and accepted practices in the electric utility industry are presented.

11.A.11 I EEE Std 1243-1997: IEEE Guide for Improving the Lightning Performance of Transmission Lines Procedures for evaluating the lightning outage rate of overhead transmission lines at voltage levels of 69 kV or higher are described. Effects of improved insulation, shielding, coupling and grounding on backflashover, and shielding failure rates of transmission lines are then discussed. Finally, a description of the IEEE FLASH program to predict lighting outage rates from www.ieee.org/pes-lightning is provided.

11.A.12 I EEE Std 1313.1-1996: IEEE Standard for Insulation Coordination—Definitions, Principles, and Rules The procedure for selection of the withstand voltages for equipment phase-to-ground and phase-tophase insulation systems is specified. A list of standard insulation levels, based on the voltage stress to which the equipment is being exposed, is also identified. This standard applies to three-phase ac systems above 1 kV.

11.A.13 I EEE Std 1410-2010: IEEE Guide for Improving the Lightning Performance of Distribution Lines Procedures for evaluating the lightning outage rate of overhead distribution lines at voltage levels below 69 kV are described. The guide then identifies factors that contribute to lightning-caused faults on the line insulation of overhead distribution lines and suggested improvements to existing and new constructions, including evaluation of the effect of wood or fiberglass materials in series with standard polymer or ceramic insulators.

References Baba, Y. and Rakov, V.A., On the Interpretation of ground reflections observed in small-scale experiments simulating lightning strikes to towers, IEEE Trans. on Electromagnetic Compatibility, 47(3), August 2005. Bewley, L.V., Traveling Waves on Transmission Systems, New York: Dover, 1963. Chisholm, W.A. and Janischewskyj, W., Lightning surge response of ground electrodes, IEEE Trans. Power Delivery, 4(2), 1329–1337, April 1989. Chow, Y.L. and Srivastava, K.D., Non-uniform electric field induced voltage calculations, Final Report for Canadian Electrical Association Contract 117 T 317, February 1988.

Transmission System Transients: Grounding

11-19

Chow, Y.L. and Yovanovich, M.M., The shape factor of the capacitance of a conductor, J. Appl. Phys., 53, 8470–8475, December 1982. Darveniza, M., A practical extension of Rusck’s formula for maximum lightning-induced voltages that accounts for ground resistivity, IEEE Trans. Power Delivery, 22(1), 605–612, January 2007. Deri, A., Tevan, G., Semlyen, A., and Castanheira, A., The complex ground return plane – A simplified model for homogeneous and multi-layer earth return, IEEE Trans. Power Appar. Syst., PAS-100(8), 3686–3693, August 1981. Grigsby, L.L., Transmission line structures (Chapter 9), Electric Power Generation, Transmission and Distribution, Electric Power Engineering Handbook, 3rd edn., CRC Press, 2011. IEEE Standard 100, The Authoritative Dictionary of IEEE Standard Terms, 7th edn., Piscataway, NJ: IEEE, 2000. IEEE Standard 80-2000, IEEE Guide for Safety in AC Substation Grounding, Piscataway, NJ: IEEE, 2000. Keller, G.G. and Frischknecht, F.C., Electrical Methods in Geophysical Prospecting, New York: Pergamon, 1982. Korsuncev, A.V., Application on the theory of similarity to calculation of impulse characteristics of concentrated electrodes, Elektrichestvo, 5, 31–35, 1958. Laurent, P.G., Les bases générales de la technique des mises à la terre dans les installations électriques, Bulletin de la Société FranÇaise des Electriciens, 1(7), 368–402, July 1951. Liew, A.C. and Darveniza, M., Dynamic model of impulse characteristics of concentrated earths, IEE Proc., 121(2), 123–135, February 1974. Loyka, S.L., A simple formula for the ground resistance calculation, IEEE Trans. EMC, 41(2), 152–154, May 1999. Oettle, E.E., A new general estimation curve for predicting the impulse impedance of concentrated earth electrodes, IEEE Trans. Power Delivery, 3(4), October 1988. Smythe, W.R., Static and Dynamic Electricity, New York: McGraw-Hill, 1950. Sunde, E.D., Earth Conduction Effects in Transmission Systems, Toronto, Ontario, Canada: Van Nostrand, 1949. Visacro, S. and Portela, C.M., Soil permittivity and conductivity behavior on frequency range of transient phenomena in electric power systems, Proc. 5th ISH, Paper 93.06, Germany: Braunschweig, August 1987. Weber, E., Electromagnetic Fields Theory and Applications Volume 1 – Mapping of Fields, New York: Wiley, 1950. Zaborsky, J., Efficiency of grounding grids with nonuniform soil, AIEE Trans., 74, 1230–1233, December 1955.

12 Transient Recovery Voltage

Thomas E. McDermott Meltran, Inc.

12.1 Fault Interruption Process............................................................. 12-1 12.2 Analysis Principles.......................................................................... 12-2 12.3 TRV for Transformer-Fed Faults...................................................12-5 12.4 TRV for Capacitor Bank Switch Opening...................................12-6 12.5 TRV for Line-Fed Faults............................................................... 12-10 12.6 TRV for Current-Limiting Reactor Faults................................. 12-13 12.7 Switchgear Tests and Standards.................................................. 12-14 12.8 TRV Mitigation.............................................................................. 12-17 References................................................................................................... 12-18

Transient recovery voltage (TRV) refers to the voltage appearing across circuit breaker contacts after it interrupts current. The circuit breaker must withstand TRV in order to complete the current interruption process. For circuit breakers rated above 1 kV, the TRV is a crucial application criterion, along with several other important factors:

1. Voltage rating 2. Interrupting current ratings 3. Capacitive current switching (inrush, outrush, close-and-latch ratings) 4. Out-of-phase switching 5. Generator breakers pose a special case due to asymmetrical current interruption

TRV is not a consideration for low-voltage circuit breakers rated 1 kV or less, but above 1 kV, every circuit breaker application should include TRV analysis. The rest of this section introduces TRV calculations and mitigation, at a level sufficient for educational purposes. For actual design work, the user should have access to one of the standard application guides (IEEE C37.011-2005, 2005; IEC 62271-100, 2008). In some cases, time-domain simulation in an electromagnetic transients (EMT) program is necessary.

12.1 Fault Interruption Process As a circuit breaker opens to interrupt current, an arc develops between the separating contacts. The current continues to flow through this arc, with a nonlinear voltage drop that produces heat. In AC steady state, the current approaches each zero crossing with approximately linear slope, and when it passes through zero, there is a chance for current interruption. During the first few microseconds after interruption, a post-arc current may flow and produce more heat. At the same time, the arc dissipates heat through radiation, convection, and conduction according to the breaker design. During this initial energy balance period, the arc must lose heat faster than it gains heat. Otherwise, a reignition occurs and interruption cannot occur until the next natural current zero crossing. Beyond the energy balance period, recovery voltage continues to build up across the still opening breaker contacts. The dielectric strength, which depends on contact separation, must exceed the 12-1

12-2

Power Systems

recovery voltage stress throughout the dielectric stress period. If not, a reignition or restrike occurs and interruption must wait for the next natural current zero crossing. The circuit breaker must pass both the energy balance and dielectric stress periods, in order to complete a current interruption. Detailed arc physics models have been developed to analyze these processes, but these models are of most use to circuit breaker designers. In applications, the TRV is usually calculated based on the electrical system model only, ignoring interactions between the arc and system models. This calculated TRV stress is then compared to standard TRV test results, which are performed in the lab and already encompass the arc behavior. There may be system TRV studies that require detailed arc models, but they are beyond the scope of this chapter. There are consequences when the breaker fails to interrupt because of TRV. At a minimum, each reignition or restrike produces a transient overvoltage, until the breaker finally interrupts. If the breaker never succeeds to interrupt, it will require inspection and repair because of thermal damage or dielectric breakdown. TRV failures usually occur when interrupting fault current, in which case a backup circuit breaker will have to interrupt the fault. Backup clearing takes longer, which allows more time for equipment damage to occur from heating or short-circuit forces during the fault. Backup clearing also removes more of the system from service, which may lead to cascading outages. If the TRV failure occurs during a normal (i.e., nonfault) opening, it probably leads to a short circuit in the failed breaker, which also leads to backup clearing and possible cascading effects. TRV analysis is important because it affects protection of the power system.

12.2 Analysis Principles Hand calculation and computer simulation both play roles in TRV studies. First, a simplified circuit analysis provides guidance on how much of the power system to model, which component parameters are most important, and what results to expect. Second, computer simulation with an EMT model should provide more realistic results and precise evaluation of all important cases. Third, reanalysis of an equivalent circuit can help explain the results and correct modeling errors. IEEE Std. C37.011 is a good source of typical data for equipment capacitance and other important parameters. Circuit breaker modeling guidelines are found in (IEEE PES Task Force, 2005). Guidelines for modeling the balance-ofsystem are found in (IEC TR 60071, 2004). TRV phenomena may consist of high-frequency transients in lumped-parameter circuits, along with traveling wave transients in distributed-parameter overhead lines and cables. It is also necessary to consider all three phases, and neutral grounding. These factors make TRV, or any transient analysis, more complicated than steady-state calculations at power frequency. Figure 12.1 shows a simplified three-phase equivalent circuit on the source side of a circuit breaker, which is interrupting a three-phase grounded fault on its terminals. The three phases do not interrupt simultaneously, and Figure 12.1 shows just the first phase opening. The most severe TRV usually appears on the first phase of the last circuit breaker to clear a fault. The source inductance often has unequal zero sequence and positive sequence values, as in Figure 12.1, which affects the equivalent inductance for TRV analysis:

Leq =

3L0 L1 L1 + 2L0

L1 (L0 – L1)/3

FIGURE 12.1 Three-phase equivalent circuit for first pole to clear.

(12.1)

C

12-3

Transient Recovery Voltage L1

L1

C

C

(Grounded)

L1/2 2C

(Ungrounded)

FIGURE 12.2 Single-phase equivalent circuits for the first pole to clear, L 0 = L1.

The equivalent capacitance, C, comes from buswork, power transformers, instrument transformers, bushings, and other equipment. This stray capacitance is usually grounded for the purpose of TRV calculations. The most important capacitance values come from power transformers, coupling capacitor voltage transformers (CCVT), and capacitive voltage transformers (CVT). Because capacitance mitigates TRV, the study should make conservative assumptions about which capacitive equipment will be in service when the breaker opens. In many cases, it is not necessary to explicitly model the buswork between equipment, but Section 12.6 describes one exception to that guideline. The left-hand side of Figure 12.2 shows the greatest possible simplification of Figure 12.1, with a grounded fault and equal sequence inductances. Each phase operates independently. This lumped circuit has transient response governed by the surge impedance and natural frequency:

L C

(12.2)

1 2π LC

(12.3)

Z lump =

f0 =

Zlump is the ratio of peak transient voltage to peak transient current in the circuit. With no damping, the transient voltage and current will oscillate around their steady-state values, with 100% overshoot. A parallel resistance in Figure 12.2 will provide damping, and reduce the peak transient voltage and current. If the parallel R ≤ 0.5 Zlump, then the circuit response will be critically damped or overdamped. This means the response is exponential rather than oscillatory. If the parallel R is approximately 4 Zlump, the damping factor is 1.7. This means the first transient peak will be only 85% of the undamped peak. The generalized damping curves in Greenwood (1991) provide more detailed information. IEEE Std. C37.011 suggests that practical damping factors range from 1.6 to 1.9, based on measurements. In an EMT model, damping comes primarily from frequency-dependent line and transformer losses, which are sometimes burdensome to model in detail. As an alternative, series or shunt resistors may be added to the EMT model at strategic points, in order to provide high-frequency TRV damping. The right-hand side of Figure 12.2 illustrates the use of circuit folding and phase symmetry, if the fault is ungrounded, but the sequence inductance values are still equal. The source-side parameters, L1 and C, do not change. However, the fault current no longer flows to ground, but must return in the two unfaulted phases (top and bottom wires in Figure 12.1). These phases provide parallel paths, with equivalent inductance L1/2 and capacitance 2C. Overhead lines and cables connected to a bus would initially appear as shunt resistors in Figures 12.1 and 12.2. Some traveling wave effects are discussed in Section 12.5. For TRV studies, the surge impedances should be calculated at high frequency, or in EMT simulations, a frequency-dependent model can be used. Typical values of the positive sequence overhead line surge impedance are 350 Ω for single conductors and 275 Ω for bundled conductors. During faults, subconductor clashing may increase the effective surge impedance. Given the positive sequence surge impedance, Z1, and the number of lines connected to a bus, the zero sequence and single-phase equivalent surge impedances are

12-4

Power Systems 1

Fault Injected

Current

0.5 0

−0.5 TRV = 0 −1

0

5

Radians

10

TRV

(Faulted) (Interruption)

FIGURE 12.3 Current injection to simulate fault interruption.

Z 0 ≈ 1.6Z1 Z eq =

3 Z 0 Z1 n Z1 + 2Z 0

(12.4) (12.5)

A cable’s surge impedance, both positive sequence and equivalent, ranges from 20 to 75 Ω. For hand calculation, the current injection method simulates TRV following fault interruption. Referring to the right-hand side of Figure 12.3, during the fault, current flows through the circuit breaker pole, and the “TRV” across that pole is 0. If the circuit is linear, then we can simulate fault interruption by injecting equal and opposite current through the breaker pole. In the left-hand side of Figure 12.3, interruption is simulated by injecting a current at π radians, and after that the sum of the two currents is 0. In the right-hand side of Figure 12.3, this is done by injecting currents on each side of the open pole. Each injected current source produces a transient voltage at its terminal, and the difference between them is the total TRV. If the fault is grounded, one of the current source terminal voltages may be 0. For hand calculation by Laplace transform, sometimes the injected currents are simplified to linear ramps, valid for a short period of time. The current injection method is also useful in EMT simulation, because the time of fault interruption is precisely controlled. That helps control numerical oscillations, and also simplifies waveform evaluation when the TRV starts exactly at time zero. The injection current sources should be sinusoids instead of ramps. Of course, the EMT model must be linear in order for superposition to apply. That is not a serious restriction because most TRV studies are done without surge arrester models and without iron core saturation models. Exceptions might apply when studying surge arresters across the breaker terminals (i.e., not to ground), and when studying lower-frequency resonant or dynamic overvoltages, for which iron core saturation may become important. The most severe TRV comes from a three-phase ungrounded fault, but those faults are quite rare. The standard TRV application for an effectively grounded system is based on a three-phase grounded fault at the breaker terminals. With EMT simulation, the question will arise about how much of the surrounding power system to model. One of the classic rules of thumb has been “within two buses from the breaker under study.” A better rule accounts for the rated TRV time to peak, which increases with the breaker voltage level, and on typical wave travel times for overhead lines. As a guideline, include all buses within 0.48 * (system nominal kV). For example, for a 115 kV system, include all buses within 55 km of the studied bus, along with all transformer and generator sources connected to those buses. This ensures that all traveling wave and remote source effects appear at the station under study, before the breaker’s rated TRV reaches its peak value.

12-5

Transient Recovery Voltage

12.3 TRV for Transformer-Fed Faults Transformer-fed faults occur quite often in medium-voltage facilities, and provide a good starting point for understanding TRV. Figure 12.4 shows a fault cleared by a breaker on the secondary of a transformer, with equivalent circuit parameters LS and CS . LS primarily comes from the transformer inductance, while CS comes primarily from transformer secondary capacitance, and secondary cables between the transformer and breaker. In this example, E is 11.3 kV peak line-to-ground (for a 13.8 kV system), LS is 0.69 mH, and CS is 0.11 μF. From Equation 12.3, the natural frequency is 18.3 kHz. The TRV will be a 1-cosine shape defined by Equation 12.6 and plotted in Figure 12.5: t TRVS = 2E 1 − cos LSCS

(12.6)

In Figure 12.6, the same breaker interrupts a fault on the other side of a load transformer, as might occur during backup clearing. The source-side equivalent circuit does not change, but there is also a load-side equivalent circuit consisting of LL and CL , where LL comes from the load transformer, and CL comes LS E

CS

FIGURE 12.4 Single-frequency transformer fed fault TRV. 25

20

kV

15

10

5

0

0

50

100

150

200

Microseconds

FIGURE 12.5 Single-frequency TRV. LS E

LL CS

FIGURE 12.6 Double-frequency transformer fed fault TRV.

CL

12-6

Power Systems 25

Source Load TRV

20

kV

15 10 5 0 –5 –10

0

50

100 Microseconds

150

200

FIGURE 12.7 Double-frequency TRV.

from the load transformer and connected cable. In this example, LL is 3.03 mH and CL is 4 nF. A doublefrequency TRV will appear across the breaker pole. The source-side component frequency is 18.3 kHz as before, while the load-side frequency is 45.7 kHz from Equation 12.3. The faulted system voltage at the breaker is no longer zero at the time of interruption, because of the load transformer impedance between breaker and fault:

Ebus = E

LL = 9.2 (kV) LS + LL

(12.7)

The load-side voltage will oscillate around zero, starting from Ebus. The source side voltage will oscillate around E = 11.3 kV, also starting from Ebus. These voltages and the TRV are provided in Equations 12.8 through 12.10 and plotted in Figure 12.7:

Eload = 9.2 cos

t LL C L

t Esource = 9.2 + 2.1 1 − cos L S CS

(12.8)

ETRV = Esource − Eload

(12.9) (12.10)

These examples illustrated single-frequency and double-frequency TRV, and the influence of impedance between breaker and fault. Damping was not considered. System resistance was also ignored, so the fault current and source voltage were 90° out of phase. Considering the actual phase shift would affect E.

12.4 TRV for Capacitor Bank Switch Opening This example uses the case of capacitor bank de-energization to illustrate the importance of neutral voltage shifts in TRV. These phenomena occur at power frequency, so the voltage across breaker contacts is more properly termed recovery voltage, rather than transient recovery voltage. However, a restrike or reignition of the breaker will produce high-frequency transients. Capacitor banks are usually switched often. Special-purpose breakers or switches may be specified to minimize the possibility of restrikes.

12-7

Transient Recovery Voltage A΄

A

FIGURE 12.8 One phase of grounded capacitor bank opening.

Ia Va΄ Va

Per-unit

1.00

0.00

−1.00 0.00

1.57

3.14

4.71

6.28 7.85 Radians

9.42

11.00 12.57

FIGURE 12.9 Grounded capacitor bank recovery voltage.

Figure 12.8 shows one phase of a three-phase grounded capacitor bank, so each phase may be considered independently, and the neutral voltage remains at zero. Figure 12.9 shows that the current leads the voltage by 90°, and when the current passes through zero at 1.5 π radians, the voltage on the bank is at its negative peak. This voltage is trapped on the disconnected capacitance, and will remain at this DC value for typically several minutes. Meanwhile, the source-side voltage continues to oscillate at power frequency. The peak voltage across the switch, Va′–Va, is 2 per unit. The switch should be able to withstand this. Figure 12.10 shows a three-phase ungrounded capacitor bank. The neutral is actually grounded through stray capacitance, which may be on the order of 1 nF, but the bank neutral voltage is no longer held at zero. The bank neutral can shift and hold a nonzero voltage, supported by the stray capacitance. The actual value of this stray capacitance has little importance to the analysis. Figure 12.11 shows the three-phase voltages on the bus, and Figure 12.12 shows the capacitor bank currents. Phase A opens first, at which time phases B and C are still energized in series, by the line-to-line voltage between phases B and C. The capacitor bank currents in phases B and C must be equal and opposite. A΄

A

B΄

B

C΄

C

FIGURE 12.10 Ungrounded capacitor bank opening.

N

12-8

Power Systems

VA΄ VB΄ VC΄

Per-unit

1.00

0.00

–1.00 0.00

1.57

3.14

4.71

6.28 7.85 Radians

9.42

11.00 12.57

FIGURE 12.11 Ungrounded capacitor bank source voltages.

IA IB IC

Per-unit

1.00

0.00

–1.00 0.00

1.57

3.14

4.71

6.28 7.85 Radians

9.42

11.00 12.57

FIGURE 12.12 Ungrounded capacitor bank opening currents.

Figure 12.12 shows these currents have equal and opposite values of 0.866 per unit, and their waveshapes change slope at the point of phase A interruption. The phase B and C currents remain equal and opposite, until they interrupt 1/4 cycle later. Figure 12.13 shows the voltage on each of the four capacitors shown in Figure 12.10. After phase A interruption, 1/4 cycle into the event, the phase A-N capacitance has a trapped DC voltage of 1 per unit. This is similar to Figure 12.9 although opposite in polarity. The phase B-N and C-N capacitances have equal −0.5 per unit voltages at this time, but they continue to be charged or discharged by the phase B-C line-to-line voltage for another 1/4 cycle. Based on Figure 12.12, phase B-N receives positive current and charge for 1/4 cycle, while phase C-N receives negative current and charge over that 1/4 cycle. The change in voltage is 0.866 per unit instead of 1.0 per unit during that time, because each capacitor is now energized by half of the line-to-line voltage, rather than line-to-neutral voltage. When phase B and C interrupt at 1/2 cycle into the event, they have trapped capacitance voltages 0.366 and 1.366 per unit, respectively.

12-9

Transient Recovery Voltage

Per-unit

1.00

0.00

VAN VBN VCN VN

–1.00

0.00

1.57

3.14

4.71

6.28 7.85 Radians

9.42

11.00 12.57

9.42

11.00 12.57

FIGURE 12.13 Ungrounded capacitor bank opening voltages. 2.00 1.50 1.00

TRVA TRVB TRVC

Per-unit

0.50 0.00 –0.50 –1.00 –1.50 –2.00 –2.50 –3.00 0.00

1.57

3.14

4.71

6.28 7.85 Radians

FIGURE 12.14 Ungrounded capacitor bank recovery voltages.

The neutral point voltage has to increase during that second 1/4 cycle of unbalanced charging, and has 0.5 per unit trapped on it when phases B and C interrupt. Figure 12.14 shows the recovery voltage on each switch pole, evaluated from TRVA = VA′ − (VAN + VN ) TRVB = VB′ − (VBN + VN )

(12.11)

TRVC = VC′ − (VCN + VN )

The peak recovery voltage on phase A is now 2.5 per unit because of the neutral shift, although the recovery voltages on the other two phases are less than 2.0 per unit. Again, the capacitor switching device should be able to withstand this. If phases B and C do not interrupt at their first chance, the neutral voltage may reach a peak of 1.0 per unit and increase the peak TRV to 3.0 per unit. If that causes a restrike, the switch should be designed to minimize the occurrence rate of restrikes.

12-10

Power Systems

The textbook (Greenwood, 1991) contains more detail on the effect of source impedance, the transient after a restrike, and the effect of neutral voltage shift on shunt reactor switching.

12.5 TRV for Line-Fed Faults When interrupting a fault fed at least partially by overhead/underground lines, traveling wave effects and resistive surge impedances play a role in the TRV, which may no longer be oscillatory (Colclaser and Buettner, 1969; Colclaser, 1972). Figure 12.15 shows an example 115 kV substation bus, fed by three transmission lines and a transformer. Two fault scenarios are considered, one at the terminals of a line breaker and another one located a short distance, d, out on the line. These two scenarios are chosen to show the difference between terminal faults and short-line faults on the same breaker. This terminal fault does not produce the highest possible fault current; the highest bus fault is fed by all three lines plus the transformer. If it is possible for a breaker within the station to interrupt the total bus fault current, that case should be included in the study because higher fault current produces higher TRV. The transformer inductance is 20 mH and the total bus capacitance is 10 nF. From Equation 12.2, the lumped circuit surge impedance is 1414 Ω. The effective surge impedance is 420 Ω for each line, and for two lines in parallel on the source side, the equivalent surge impedance Z = 210 Ω. This value is much less than half of the lumped circuit surge impedance, so the TRV will be an exponential waveform. The capacitance is small enough to be ignored for initial analysis. If the fault current through the line breaker is 20 kA, then the TRV response is given in Equation 12.12, where the angular frequency ω is 377 for a 60 Hz power system (i.e., this case) and 314 for a 50 Hz power system: τ=

L = 95 (µs) Z

E1 = 2 I flt ωL = 213 (kV)

(12.12)

t TRV = E1 1 − exp τ

This TRV is plotted in Figure 12.16. In a real power system, traveling wave reflections from nearby stations would modify this TRV, possibly increasing or decreasing the peak value. It is difficult to analyze these reflections by hand, because they are complicated by multiple line and inductive terminations. On a 115 kV system, about 9 kA fault current would flow through the transformer, and the remaining 11 kA must come from the two unfaulted lines. If the fault current were higher, then the TRV would increase. On the other hand, higher fault current usually means more parallel source connections, and stronger source equivalents at nearby stations, both of which should help reduce the TRV. This is one of the main reasons for using EMT simulation to perform the study. The EMT model can be set up for

d

FIGURE 12.15 Bus terminal and short-line fault fed by lines and transformers.

12-11

Transient Recovery Voltage 225 200 175

kV

150 125 100 75 50 25 0

0

50

100

150 200 Microseconds

250

300

FIGURE 12.16 Bus terminal fault TRV.

transients only, and use current injection sources derived from a separate fault study. Alternatively, the EMT model can be set up and used for both steady-state fault analysis and transient analysis, avoiding the use of current injection. Either way, accurate fault current levels must be used in the study. If there is no inductive source at the station in Figure 12.15, then the TRV will consist of a linearly rising ramp, until modified by traveling wave reflections from nearby stations. Cables have lower surge impedances than overhead lines, leading to longer time constants for exponential TRV, or lower slopes for linear TRV components. The second fault location in Figure 12.15 is a distance d = 3.2 km from the line breaker terminals. The TRV is lower because the fault current is less, but a saw tooth component on the 3.2 km line segment produces a higher initial rate of rise. Years ago, short-line faults like this led to some breaker failures. Newer versions of the IEEE and IEC standards include short-line fault tests, so that the user does not have to explicitly evaluate them for applications. However, the case provides an instructive example, and the tested short-line fault capability can also be applied to other situations. Assuming the maximum fault currents were evaluated at 1.05 per unit operating voltage, the first step is to estimate the breaker fault current during a short-line fault. (This could also be done using a separate short-circuit analysis.) Based on typical line reactance of 0.5 Ω/km in positive sequence and 1.2 Ω/km in zero sequence, the total line reactance to the fault is

XL = d

2 X1L + X0 L = 2.35 Ω 3

(12.13)

The source-side equivalent reactance, from two lines and one transformer, comes from the 20 kA terminal fault current:

XS =

1.05 * 115 = 3.49 Ω 3 * 20

(12.14)

Those two reactances now determine the reduced fault current:

I SLF =

1.05 * 115 = 11.9 kA 3 * (2.35 + 3.49)

(12.15)

12-12

Power Systems

The bus voltage at the time of interruption is no longer zero, because of the line impedance out to the fault. After interruption, the source-side voltage will approach its 1.05 per unit peak prefault level according to an exponential waveform. Modifying Equation 12.12 accordingly, τ=

L = 95 (µs) Z

2 − X L I SLF 2 = 59 (kV) E1− SLF = 115 * 1.05 * 3

(12.16)

t TRV1− SLF = E1− SLF 1 − exp τ

The source-side TRV component, plotted in Figure 12.17, has a lower prospective peak value but the same time constant. Equation 12.17 defines the rate of rise of the line-side component; Zeff = 420 Ω because it applies to the single faulted line segment R = 2 ωI SLF Zeff * 10 −6 = 2.69(kV/µs)

(12.17)

In IEEE standards, a damping factor of 1.6 applies to the line-side component of TRV, meaning that the damped peak reaches 1.6 per unit of the line-side breaker terminal voltage at the instant of interruption. Without damping, the peak would be 2.0 per unit of that line-side breaker terminal voltage. This damping factor is η, and the peak line-side TRV component is U L = ηX L I SLF 2 = 63.3 kV

75

(12.18)

Source side Line side

50

kV

25 0 –25 –50 –75

0

20

40

FIGURE 12.17 Short-line fault TRV components.

60 80 Microseconds

100

120

140

12-13

Transient Recovery Voltage

The time to first peak is UL/R = 23.5 μs. Figure 12.17 includes this damped line-side TRV component. The total TRV is the difference between the two waveforms in Figure 12.17. Compared to Figure 12.16, the peak TRV is lower but the initial rate of rise is more severe (note that for clarity, Figure 12.17 has a shorter timescale than Figure 12.16).

12.6 TRV for Current-Limiting Reactor Faults Shunt capacitor banks and underground cables often have current-limiting reactors (CLR) installed in series. The CLR may be necessary to limit capacitor bank inrush and outrush currents, especially when multiple capacitors are connected to the same bus. CLR has also been used to reduce fault currents in low-impedance cable systems. The CLR can increase high-frequency TRV stresses on switchgear, and there have been cases of capacitor switch failure when trying to interrupt a CLR-limited fault. When this causes backup clearing of several lines and transformers connected to that bus, more widespread outages and disruptions can occur. Figure 12.18 shows a simplified view of one station with a faulted capacitor bank at the node CAP, and a breaker at node BRKR tasked to clear that fault. A CLR between the capacitor and breaker has a series inductance, with relatively small shunt and series capacitances associated. The high-frequency lumped-circuit oscillations within the CLR produce high TRV rate of rise on the line side, and possibly lead to failure. Because of the fault location, the capacitor bank’s large capacitance provides no TRV mitigation. However, other buswork and equipment capacitance located between CLR and breaker may help in mitigating the TRV. On the source side of the breaker, a total bus capacitance and external network of lines and transformers determine the source-side TRV, which would be similar to that produced by a bus terminal fault. If the CLR were moved to the source side of the breaker, it would still produce a high-frequency TRV component. That high-frequency TRV, now on the source side instead of load side, would still possibly lead to breaker failure. Figure 12.19 shows a simplification of the model, by aggregating all capacitance between breaker and CLR into one value, ignoring the inductance of connections between them. The full external system network should still be part of the model, for accurate fault currents and source-side TRV components. The study can focus on changing CLR or breaker parameters to mitigate the TRV. CCLR

CAP

D4

CLR

LCLR

D1 CBANK Fault

150 pF

CVT

TAP

BANK

150 pF

D3 D2 BRKR SYSBUS

CBUS

FIGURE 12.18 Model for current-limiting reactor fault TRV.

CCVT System network

12-14

Power Systems CCLR LCLR

BRKR

SYSBUS

150 pF + CCVT + C΄(D1 + D2 + D3)

CBUS

System network

Fault

FIGURE 12.19 Reduced model for current-limiting reactor fault TRV.

12.7 Switchgear Tests and Standards Both IEEE and IEC define circuit breaker test procedures, rating structures, preferred rating values, and TRV evaluation procedures in a set of standards that are updated periodically (IEEE C37, IEC 62271). Most North American utilities use IEEE standards, while other utilities in the world generally use IEC standards. The same choice would apply to medium-voltage and high-voltage industrial or commercial users, who might also need to do TRV studies. Historically, there have been differences between IEEE and IEC standards, but the IEEE standards have been recently harmonized with IEC. That harmonization process is not yet complete. Even when harmonization is completed, circuit vendors can still offer products that were type-tested to earlier versions of IEEE standards. Circuit breakers can last for many decades in service, and they would have been tested to earlier standards. For both IEEE and IEC standards, there may be changes in procedures and ratings between updates. These are all good reasons to verify, at the beginning of a TRV study, which version of the standards applies to the breakers under evaluation. At present, the standards mention that TRV evaluation can be based on three-phase grounded terminal faults, and that the vendor is responsible for testing breakers to withstand short-line fault TRV. There may be situations that still call for analysis of three-phase ungrounded faults, short-line faults, out-of-phase switching, or other examples not covered in this chapter. Figure 12.20 shows the TRV rating for a 123 kV class circuit breaker, interrupting 100%, 60%, 30%, or 10% of its rated fault capability. These ratings apply to older IEEE standards; at 60% and TRV capability for 123 kV breaker; pre-2009 IEEE

225 200 175 150 kV

125 100 75 50

M = 100% M = 60% M = 30% M = 10%

25 0

0

50

FIGURE 12.20 Old IEEE TRV ratings.

100

150 Microseconds

200

250

300

12-15

Transient Recovery Voltage TABLE 12.1 Interpolated TRV Ratings for IEEE and IEC Standards, Breakers Rated 100 kV and Higher IEEE

IEC

M (%)

kE2

kT2

R1 (kV/μs)

Kaf

ROR (kV/μs)

T2

10 30 60 100

1.17 1.13 1.07 1.00

0.2 0.2 0.5 1.0

0 0 4 2

1.53 1.54 1.50 1.40

7 5 3 2

0 Uc/ROR 6T1 4T1

higher, the TRV envelope is a composite 1-exponential and 1-cosine shape. At 30% and 10% interrupting rating, and for voltage ratings below 100 kV, only the 1-cosine shape is used. At fault current levels in between the defined four test levels, TRV rating envelopes may be interpolated. The calculated or simulated TRV stress should lie entirely under the applicable envelope in Figure 12.20 for a successful application. The user should also be aware of the TRV “delay line” as described in both IEEE and IEC standards. Given breaker rating values of Em = 123 kV and T2 = 260 μs, the curves in Figure 12.20 may be constructed from Equation 12.19 with interpolation factors from Table 12.1.

K a = 1. 4 K f = 1. 3 T2 ← T2kT2 U r = Em

2 3

E2 = K a K f U r kE2 τ=

E1 R1

(12.19)

t Eexp = E1 1 − exp − τ πt Ecos = 0.5E2 1 − cos T2

TRV = max(Eexp , Ecos )

R1 is a rate of rise, to be interpolated from Table 12.1. If R1 is 0, the Eexp term does not appear in the TRV, only the 1-cosine term, Ecos. Table 12.1 also provides multipliers to use with E2 and T2, at lower-thanrated fault currents. Figure 12.21 shows corresponding IEC standard TRV ratings, for the same four test current levels, and the same 123 kV breaker voltage class. At 100% and 60% of rated interrupting current, the TRV envelope is a two-slope characteristic, defined by two points, or four parameters. At 30% and 10% of rated interrupting current, or for any breaker voltage rating below 100 kV, only one point defines the TRV. These are often called “4 parameter” and “2 parameter” TRV ratings. Newer versions of the IEEE

12-16

Power Systems TRV capability for 123 kV breaker; IEC

225 200 175 150

kV

125 100 75 50

M = 100% M = 60% M = 30% M = 10%

25 0

0

50

100

150 Microseconds

200

250

300

FIGURE 12.21 IEC and harmonized IEEE TRV ratings.

standards will use TRV ratings presented as in Figure 12.21. With reference to interpolated values from Table 12.1, Figure 12.21 is constructed according to K pp = 1.3 U r = Em

2 3

(12.20)

U c = K pp K af U r

If M ≤ 30% then Equation 12.21 applies Uc ROR T1 = 0.5T2

T2 =

(12.21)

U1 = 0.5U c

If M > 30% then Equation 12.22 applies, and T2 is interpolated according to Table 12.1 after T1 has been determined: U1 = 0.75K ppU r

T1 =

U1 ROR

(12.22)

The TRV is then constructed by linear interpolation between four points: (0, 0), (T1, U1), (T2, Uc), and (2T2, Uc). Equations 12.20 through 12.22 smoothly transition from two-parameter and four-parameter characteristics, when M increases from 30% to 60%.

12-17

Transient Recovery Voltage Composite TRV capability for 123 kV, 40 kA breaker; pre-2009 IEEE

100 90 80 70

kV

60 50 40 30 20 M = 90% M = 75%

10 0

0

5

10

15

20 25 30 Microseconds

35

40

45

50

FIGURE 12.22 Composite TRV rating based on short-line fault tests.

Figure 12.22 illustrates the aggregation of a breaker’s tested short-line fault TRV capability, with the base TRV rating. By IEEE standards, the short-line fault test may be performed at M values of 70%–95%. The standard test surge impedance Z = 450 Ω, and this breaker has rated frequency f = 60 Hz and rated interrupting current Ir = 40 kA. The short-line fault test verifies TRV withstand up to a point defined by the linear segment from (0, 0) to (Tslf, Eslf), as obtained from Eline = 1.6(1 − M )U r Rslf = 8.8858 −6 fMI r Z Tslf =

Eline Rslf

(12.23)

Esrce = max 0, 2 M (Tslf − 2) Eslf = Eline + Esrce

After Tslf, the TRV rating extends horizontally at voltage Eslf, until it encounters the exponential TRV characteristic. The extra short-time TRV capability may be useful in applications that produce a fast-rising TRV.

12.8 TRV Mitigation If the calculated TRV stress exceeds the breaker ratings, mitigation will be necessary:

1. Consult the circuit breaker vendor, and provide the TRV stress waveform in electronic form. Sometimes the vendor has designed the breaker to withstand higher TRV, and is willing to state that the application is acceptable. 2. Add capacitance to the circuit breaker terminals, or across the terminals. This slows down the TRV by reducing the rate of rise or oscillation frequency. It is one of the most common TRV mitigation techniques. However, the first peak in TRV can increase slightly with more capacitance, so that should be evaluated, too. The capacitance also has to be located close enough to the circuit breaker, and cannot be separated from the breaker through disconnect switch operation or station reconfiguration.

12-18

Power Systems TRV for 14 kA fault at IPS2 bus T100 Rating

30

TRV (1 nF)

TRV (40 nF)

T70 Rating

Voltage (kV)

25 20 15 10 5 0

0

10

20

30

40 Time (µs)

50

60

70

80

FIGURE 12.23 Mitigation of medium-voltage breaker TRV. (Reproduced with permission from McDermott, T.E. and Dafis, C., Cost-effective ship electrical system simulation, ISS IX Conference Paper. Philadelphia, PA, May 25–26, 2011. Copyright 2011 American Society of Naval Engineers.)

3. When a current-limiting reactor is involved, capacitance can be added on the CLR terminals. It may also be possible for the CLR vendor to alter its design to have more internal capacitance. 4. Use a circuit breaker of higher voltage rating. This can be expensive, but it works because the breaker will have significantly higher TRV capability for the same fault current. 5. Use a circuit breaker of higher current rating. This can also be expensive, but it increases the TRV capability because the system fault current will be a lower portion of the breaker rating. 6. Reduce the system fault current, either by splitting a bus or by adding fault current-limiting impedances. The system impedance changes may tend to increase TRV, but that is usually more than offset by the reduction in fault current level. 7. Surge arresters across the breaker contacts have been considered, especially when the TRV comes from dynamic overvoltages. One must also check the arrester energy and temporary overvoltage ratings in this application. 8. Some older circuit breakers have used grading resistors to limit TRV.

Any such measures that change the system response (e.g., items 2, 3, 6, 7, 8) call for a reevaluation of the TRV, with countermeasures in place. Figure 12.23 shows the TRV stress on an example 15 kV, 20 kA circuit breaker interrupting a 14 kA fault (Mc Dermott and Dafis, 2011). With 1 nF total bus capacitance, the TRV exceeds the breaker capability at 100% current interruptions, T100. By increasing the total capacitance to 40 nF, the peak TRV occurs later, and the stress waveform lies completely under the T100 rating. Because the 14 kA fault current is only 70% of the breaker’s interrupting capability, the TRV rating may be increased to T70. In that case, the TRV stress waveform lies completely under the T70 rating, even with 1 nF total capacitance.

References ANSI Std. C37.06-2009, IEEE Standard for AC High-Voltage Circuit Breakers Rated on a Symmetrical Current Basis–Preferred Ratings and Related Required Capabilities for Voltages above 1000 V, The Institute of Electrical and Electronic Engineers, New York, 2009. Colclaser, R. G., The transient recovery voltage application of power circuit breakers, IEEE Transactions on Power Apparatus and Systems, 91(5), 1941–1947, May 1972.

Transient Recovery Voltage

12-19

Colclaser, R. G. and D. E. Buettner, The traveling-wave approach to transient recovery voltage, IEEE Transactions on Power Apparatus and Systems, 88(7), 1028–1035, July 1969. Greenwood, A. N., Electrical Transients in Power Systems, 2nd edn., John Wiley & Sons, New York, 1991. IEC, Insulation co-ordination–Part 4: Computational guide to insulation co-ordination and modelling of electrical networks, Technical Report TR 60071-4, Edition 1, International Electrotechnical Commission, Geneva, Switzerland, June 2004. IEC Std. 62271-1, High-voltage Switchgear and Controlgear—Part 1: Common Specifications, Edition 1.0, International Electrotechnical Commission, Geneva, Switzerland, October 2007. IEC Std. 62271-100, High-voltage Switchgear and Controlgear—Part 100: Alternating-current Circuitbreakers, Edition 2.0, International Electrotechnical Commission, Geneva, Switzerland, April 2008. IEEE PES Task Force on Data for Modeling System Transients, Parameter determination for modeling system transients—Part VI: Circuit breakers, IEEE Transactions on Power Delivery, 20(3), 2079–2085, July 2005. IEEE Std. C37.04-1999, IEEE Standard Rating Structure for AC High-Voltage Circuit Breakers, The Institute of Electrical and Electronic Engineers, New York, June 1999. IEEE Std. C37.04-1999/Cor 1-2009, IEEE Standard for Rating Structure for AC High-Voltage Circuit Breakers Corrigendum 1, The Institute of Electrical and Electronic Engineers, New York, 2009. IEEE Std. C37.04a-2003, IEEE Standard Rating Structure for AC High-Voltage Circuit Breakers Rated on a Symmetrical Current Basis: Amendment 1 Capacitance Current Switching, The Institute of Electrical and Electronic Engineers, New York, July 2003. IEEE Std. C37.04b-2008, IEEE Standard for Rating Structure for AC High-Voltage Circuit Breakers Rated on a Symmetrical Current Basis Amendment 2: To Change the Description of Transient Recovery Voltage for Harmonization with IEC 62271-100, The Institute of Electrical and Electronic Engineers, New York, 2008. IEEE Std. C37.09-1999, IEEE Standard Test Procedure for AC High-Voltage Circuit Breakers Rated on a Symmetrical Current Basis, The Institute of Electrical and Electronic Engineers, New York, January 2000. IEEE Std. C37.09-1999/Cor 1-2007, IEEE Standard Test Procedure for AC High-Voltage Circuit Breakers Rated on a Symmetrical Current Basis–Corrigendum 1, The Institute of Electrical and Electronic Engineers, New York, September 2007. IEEE Std. C37.09a-2005, IEEE Standard Test Procedure for AC High-Voltage Circuit Breakers Rated on a Symmetrical Current Basis: Amendment 1, Capacitance Current Switching, The Institute of Electrical and Electronic Engineers, New York, March 2005. IEEE Std. C37.011-2005, IEEE Application Guide for Transient Recovery Voltage for AC High-Voltage Circuit Breakers, The Institute of Electrical and Electronic Engineers, New York, September 2005.

13 Surge Arresters

Thomas E. McDermott Meltran, Inc.

13.1 Arrester Types and Auxiliary Equipment................................... 13-1 13.2 Ratings and Tests............................................................................. 13-3 13.3 Selection by TOV.............................................................................13-6 13.4 Selection by Energy Rating............................................................ 13-7 13.5 Arrester Modeling...........................................................................13-9 13.6 Applications.................................................................................... 13-12 References................................................................................................... 13-16

Surge arresters are shunt, nonlinear, resistive devices that are installed to limit transient overvoltages, thereby protecting equipment insulation. Sometimes they have been called “lightning arresters,” and this may be due to the fact that many years ago, surge arresters were tailored to protect against the surge front times typical of lightning. They were much less effective against steeper-fronted overvoltages, or against slow-front switching surges. However, the proper term is “surge arresters,” and the modern generation of metal oxide surge arresters is effective against both lightning and switching surges. The rest of this chapter summarizes metal oxide surge arrester ratings and application principles for system voltage levels above 1 kV. This treatment should suffice for educational purposes. For actual design work, the use of time-domain simulation in an electromagnetic transients program (EMTP) would be highly recommended. The designer should also have access to one of the standard application guides (IEEE C62.22, IEC 60099-5) or the text on insulation coordination by Hileman (1999).

13.1 Arrester Types and Auxiliary Equipment Nearly 100 years ago, electrode gaps (rod, sphere, or pipe) were used to limit overvoltages on equipment (Sakshaug, 1991). Some of these systems, particularly pipe gaps, may still be in service today. However, the characteristic of gap sparkover voltage vs. surge front time does not match up well with the strength vs. front characteristics of most insulation; that is, it is difficult to coordinate. The next evolutionary step was to add a resistive element in series with the gap, in order to limit the power follow current after an arrester discharge operation. The current limiting would hopefully allow the arrester to clear this power follow current, instead of relying on a nearby breaker or fuse. At the same time, the resistor voltage during a discharge must be low enough that it does not allow an excessive voltage to appear on the protected equipment. These competing requirements led to the use of expensive and complicated nonlinear resistive elements, some involving both solid and liquid materials with high maintenance burdens. Beginning around 1930, silicon carbide (SiC) was used for the nonlinear resistive elements, leading to much better protective characteristics. Because the SiC would conduct significant current at nominal voltage, it was necessary to provide a sparkover gap that prevents conduction at nominal voltage. After an arrester discharge, these gaps must reseal against the power follow current, otherwise, 13-1

13-2

Power Systems

the arrester would fail thermally. In the mid 1950s, active gaps were developed for SiC arresters. These active gaps contain auxiliary elements that would

1. Preionize the sparkover gap to obtain better surge protective levels 2. Elongate the power follow arc, and move its attachment points, to obtain better interruption performance

SiC arresters were successfully applied on transmission systems up to 345 kV, but some limitations appeared with regard to switching surge protection, energy discharge capability, and pressure relief capability. Having both gaps and SiC blocks, the arrester height increased to the point where it was difficult to vent the pressure buildup during a fault, which limited the arrester’s pressure relief rating. Due to their discharge characteristics vs. frequency or front time, the SiC surge arresters were optimized for lightning. They were less effective for steeper-fronted surges and slow-fronted switching surges. In the mid 1970s, metal oxide surge arresters were developed into commercial products (Sakshaug et al., 1977). The metal oxide blocks are much more nonlinear than silicon carbide, so that they conduct only a few milliamperes at nominal AC voltage. It eventually became possible to dispense with gaps completely, although earlier designs made some use of gaps (see Figure 13.1). Metal oxide surge arresters have several major advantages over the earlier silicon carbide arresters:

1. Active gaps are not necessary, leading to improved reliability. 2. Metal oxide can discharge much more energy per unit volume than silicon carbide. 3. Metal oxide provides better protection across the range of surge wave fronts than silicon carbide, and in fact, protects effectively against switching surges. 4. The decrease in arrester height, caused by eliminating sparkover gaps, leads to higher pressure relief ratings.

Virtually all new applications will use metal oxide surge arresters. Metal oxide has enabled some new applications, like series capacitor protection and overhead line switching surge control that were not possible with silicon carbide. However, many silicon carbide arresters are still in service. Some investigators have noted high silicon carbide arrester failure rates, due to moisture ingress, after several years of service on medium-voltage distribution systems. This experience does not necessarily apply to surge arresters in substations. If such problems arise, it would make sense to systematically replace silicon carbide arresters on a system. Otherwise, assuming that the original application was proper, the older silicon carbide arresters could remain in service. Figure 13.1 shows the general use of gaps in surge arresters. The gapped design (Figure 13.1a) applies to silicon carbide, whereas the gapless design (Figure 13.1d) applies to the latest generation of metal oxide. One manufacturer used the shunt gap (Figure 13.1b) in early metal oxide arresters. At steady state, both

(a)

(b)

(c)

(d)

FIGURE 13.1 Use of gaps in metal oxide surge arresters (a) gapped, (b) shunt gap, (c) graded gap, and (d) gapless.

Surge Arresters

13-3

nonlinear elements would support the nominal voltage, somewhat reducing the current. During a surge discharge, the shunt gap would sparkover to bypass the smaller section of metal oxide, thereby reducing the discharge voltage and providing somewhat better protection. Another manufacturer used the series gap with capacitive grading (Figure 13.1c) in early metal oxide arresters. At steady state, this decreases the voltage on the metal oxide. During a surge discharge, the gap sparks over “immediately” due to the capacitive grading. The latest generations of metal oxide do not need these gaps, although there is some consideration for using gaps to achieve specific goals (e.g., coordinating arresters, withstanding temporary overvoltages). Recently, surge arresters have been applied to transmission lines for lightning protection of the line insulation. Two varieties have been developed, first a nongapped line arrester (NGLA) as in Figure 13.1d, and then an externally gapped line arrester (EGLA) as in Figure 13.1a. The surge arrester must be installed on something, such as a transformer tank or a pedestal. It must also be connected to the protected system, typically through a wire or lead. Later, it will be shown that these connections have important effects on the overall protection, especially for steep surges. The pedestal and lead, both length and location, must be considered as part of the overall arrester installation. On distribution systems, a ground lead disconnector is often used with the surge arrester. If the arrester fails and then conducts current on a steady-state basis, the disconnector will detonate and disconnect the base of the arrester from ground. This should happen in approximately 1 s, or faster. The arrester may then remain connected to the system until maintenance personnel have a chance to replace it. No breaker or fuse need operate to isolate the failed arrester; if the arrester is the only thing that failed, no customers need to lose their electric service. Of course, the arrester is not providing any surge protection during this period with its ground lead disconnected. There would be a clear visual indication that the ground lead has been disconnected; it will be “hanging down” below the arrester. Regular visual inspections are necessary to maintain surge protection whenever ground-lead disconnectors are used. Many surge arresters in substations or industrial facilities have been installed with “surge counters.” These are accessories to be installed in the surge arrester’s ground-lead connection. Two functions may be provided:

1. A steady-state current meter, calibrated in mA. If this current increases over time, it may indicate thermal damage to the surge arrester. However, the presence of harmonics or external leakage currents would complicate the assessment. 2. A counter indicates the number of surge current discharges above a certain threshold, which may depend on frequency or front time. Even if the count is accurate, it does not mean that the discharge voltage reached any particular level during those events.

To use surge counters effectively, it is important to track the readings on a regular basis, beginning at the time of commissioning.

13.2 Ratings and Tests Surge arresters have both a voltage rating and a class, or type. For metal oxide, the important voltage rating is the maximum continuous operating voltage (MCOV), which is the steady-state voltage the arrester could support indefinitely. This rating is most important for metal oxide, because most of these arresters are now gapless and carry a few milliamperes at all times. The MCOV should be at least 1.05 times the system’s nominal line-to-ground voltage. There are some cases like distribution feeders with poor voltage regulation, which might require a higher MCOV. As discussed later, short-term temporary overvoltages (TOV) also play an important role in selecting the arrester rating, but still the basic rating is MCOV.

13-4

Power Systems TABLE 13.1 Schedule of IEC Voltage Rating Steps Range of Rated Voltage (kV rms) 3–30 30–54 54–96 96–288 288–396 396–756

Step Size (kV rms) 1 3 6 12 18 24

Historically, the SiC arresters had a voltage rating that corresponded to the duty cycle test, with no direct link to MCOV. Given that SiC arresters were always gapped, MCOV was not an important concern. In IEEE standards, the old SiC numerical schedule of voltage ratings has been carried over to metal oxide arresters, and the MCOV is usually about 84% of the arrester voltage rating. For example, a typical 108 kV arrester has an MCOV of 84 kV. Only the MCOV number is important. The IEC standard defines the voltage rating to be the TOV capability at 10 s, and again the MCOV is about 84% of this value. Table 13.1 shows the schedule of IEC voltage rating steps, and in general these line up with the schedule of IEEE voltage ratings. This table could be used to choose the preliminary arrester rating for some application. However, all surge arrester manufacturers now have their catalog information readily available on the web, and it is better to use their actual data. The IEEE specifies a set of withstand tests for surge arresters, each of them followed by a period of operation at MCOV. These tests are as follows:

1. High current, short duration: Two 4 × 10 μs current surges are discharged through the arrester. The peak current is 65 kA for station, intermediate, and distribution normal-duty arresters. It is 100 kA for distribution heavy-duty and 40 kA for distribution light-duty arresters. This is followed by 30 min at MCOV. 2. Low current, long duration: For station and intermediate arresters, a transmission pi section is discharged 20 times through the arrester. The switching surge voltage ranges from 2.0 to 2.6 pu and the equivalent line length ranges from 160 to 320 km, depending on the arrester class and voltage rating. For distribution arresters, a 2000 ms square wave of current is discharged 20 times, with a peak current of 250 A for heavy-duty and 75 A for normal-duty or light-duty arresters, followed by 30 min at MCOV. 3. Duty cycle test: While energized at the voltage rating, which is higher than MCOV, discharge an 8 × 20 μs lightning impulse 20 times, with peak ranging from 5 to 20 kA, followed by 2 discharges without power frequency voltage, followed by 30 min at MCOV. 4. Pressure relief test: An AC current of 40–80 kA rms (root mean square) is applied for station class, or 16.1 kA for intermediate class, so that the arrester fails. The pressure must vent well enough that any components lie no farther away than the arrester height. Polymer housings tend to have better pressure relief ratings than porcelain housings. 5. Contamination test: Energized for 1 h at MCOV under contaminated conditions, followed by 30 min at MCOV. 6. Temporary overvoltages: Verifies the TOV capability over a range from 0.02 to 1000 s. 7. Switching surge energy: This is an optional supplement to the low-current, long-duration test, which carries an implied energy discharge duty. The manufacturer typically provides a value for energy capability in kJ per kV of MCOV, which is meant to cover multiple discharges within a period of about 1 min.

The test results for TOV and switching surge energy capability are often used in application studies.

13-5

Surge Arresters

The manufacturer also conducts tests of the arrester protective characteristics. These are always directly useful in modeling and application studies:

1. 8 × 20 μs discharge voltage: Current surges are applied with peak magnitudes of 1.5, 3, 5, 10, and 20 kA, with the resulting peak discharge voltages tabulated. Due to various frequency-dependent effects, the peak voltage occurs before the peak current. For some ratings, the 15 and 40 kA discharge voltages are also tabulated. 2. Front-of-wave (FOW) protective level: This is the voltage having a 0.5 μs time to crest, for a current surge of 5, 10, 15, or 20 kA depending on the rating. To get this value, it is necessary to apply different front times for the current surge and interpolate or extrapolate the results. 3. Switching surge protective level: The peak arrester voltage for a current surge having 45–60 μs front time and a peak of 500–2000 A depending on the rating.

Voltage (pu of V10)

Many vendors provide more detailed information about protective levels as a function of the surge front time, which is helpful in modeling. Figure 13.2 shows typical maximum 8 × 20 μs discharge voltage and FOW discharge characteristics for station class metal oxide surge arresters, rated 54 kV and above. There is a manufacturing variation within each product line, and the minimum discharge voltage could be a few percent lower than these values. In most cases, the maximum values would be used for application studies. Figure 13.3 shows

1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5

Generic lightning discharge characteristics

FOW 8 × 20 0

10

20 Current (kA)

30

40

Voltage (pu of V10)

FIGURE 13.2 Metal oxide surge arrester lightning discharge characteristics, for voltage ratings from 54 to 360 kV. V10 is the arrester’s 8 × 20 μs discharge voltage at 10 kA.

1 0.95 0.9 0.85 0.8 0.75 0.7 0.65 0.6 0.55 0.5

Generic switching discharge characteristic

0

200

400

600

800 1000 1200 Current (A)

1400

1600

1800 2000

FIGURE 13.3 Metal oxide surge arrester switching discharge characteristic, for voltage ratings from 54 to 360 kV. V10 is the arrester’s 8 × 20 μs discharge voltage at 10 kA.

13-6

Power Systems

a typical maximum switching surge discharge characteristic. All of these are normalized to the published 8 × 20 μs discharge voltage at 10 kA. The IEEE standards define several classes of surge arresters, based on the withstand test levels discussed earlier. In substations and many transmission line applications, the station class arresters are most common. The intermediate class arresters are used for lightning protection of transmission line towers and for some of the smaller substations. The choice of distribution arrester class depends mainly on how severe the local lightning environment is. Although purchased in significant quantities, the surge arresters are typically not a significant cost item in a single substation. Among the IEEE standards, C62.11 defines the metal oxide arrester tests and ratings, whereas C62.22 provides the application guidance. The older standards for silicon carbide arresters, C62.1 and C62.2, are still available. These apply to the United States, whereas, in most other parts of the world, the International Electrotechnical Committee (IEC) standards are used. These are denoted IEC 60099 parts 1 through 8, except that there is no part 2 and part 7 is a glossary. For gapless metal oxide arresters, part 4 defines the tests and ratings, while part 5 presents the application guidelines. For gapped arresters, including silicon carbide, part 1 applies. The IEEE standards may be found through www. ieee.org, and the IEC standards through www.iec.ch. Both IEEE and IEC standards define the discharge characteristics in similar ways. The same applies to MCOV and TOV capabilities. In IEC standards, the rated voltage is actually the TOV capability at 10 s and also the duty cycle test voltage. These differences generally will not concern the user. The main difference is that IEC standards classify arresters by nominal discharge current (1.5, 2.5, 5, 10, or 20 kA) and line discharge class (1–5, or none). An IEEE station class arrester is roughly comparable to an IEC 10 kA arrester, whereas the intermediate and distribution class arresters are roughly comparable to IEC 5 kA arresters. At 500 kV and above, the IEEE classifying current is 15 or 20 kA, roughly corresponding to the IEC nominal discharge current of 20 kA used at these voltage levels. The appropriate standard should be consulted for details, but Osterhout (1992) and Hamel and St. Jean (1992) provide overviews of the differences. Low-voltage surge arrester applications follow a different set of standards and guides, for system voltage levels below 1 kV, which includes secondary services and customer utilization circuits. The applicable IEEE standards are C62.72 and C62.41 for application guidance, plus others in the C62 family for test requirements. The applicable IEC standards for low-voltage power systems are 61643, parts 1 and 12.

13.3 Selection by TOV As discussed earlier, the first choice of arrester voltage rating is based on MCOV. That would be the lowest and best choice because it minimizes the arrester discharge voltage for any particular surge, thereby maximizing protective margins. However, it may be necessary to increase the arrester voltage rating to withstand either TOV or surge energy content. The TOVs come from several sources, most commonly

1. During ground faults, voltage on the unfaulted phases will rise above nominal unless the system is solidly grounded. The duration of this overvoltage depends on protective relaying system response, usually less than 1 s, but up to several hours for delta systems or ungrounded systems. 2. When first energizing a long line or cable, the voltage at the open end will rise above nominal due to the Ferranti effect. The duration depends on how long it takes to switch on shunt compensation or to close the other end of the line. A similar effect occurs after load rejection when only one end of the line opens. 3. Harmonics, ferroresonance, and transformer inrush currents can produce dynamic overvoltages that last for several cycles.

Usually the ground fault TOVs are most important, but there are special cases where simulation of other TOVs should be considered. Given the zero sequence and positive sequence impedances at the point of

13-7

Surge Arresters Temporary overvoltage capability

Per unit of MCOV

1.7

Distribution class No prior energy With prior energy

1.6 1.5 1.4 1.3 1.2 1.1 0.01

0.1

1

10 Time (s)

100

1000

10000

FIGURE 13.4 Temporary overvoltage capability.

a fault, the per-unit unfaulted phase voltage, often called the earth-fault factor (EFF) may be estimated for a single-line-to-ground fault (IEEE C62.22) EFF = − K=

3 2

3K 2 + K ± j1

Z 0 + Rf Z1 + Rf

(13.1)

Z 0 = R0 + jX0

Z1 = R1 + jX1

Note that there are two unfaulted phases to evaluate, corresponding to the +j1 and −j1 terms. That voltage should be further increased by a factor of 1.05 to account for above-nominal operating voltage. In most cases, the fault resistance, Rf, should be considered zero as it maximizes EFF. For distribution systems, Rf should be equal to the design level of fault resistance that relaying will detect; this is really a matter of policy. For double-line-to-ground faults, the per-unit voltage on the unfaulted phase would be (IEEE C62.22) EFF =

K=

3K 2K + 1 Z 0 + 2Rf Z1 + 2Rf

(13.2)

Again, this voltage should be further increased by a factor of 1.05 to account for above-nominal prefault voltage. The TOV duties from various sources must be compared to the arrester’s TOV capability, which may be obtained from the manufacturer’s catalog, or taken from Figure 13.4 to use typical data. Some data include TOV capability both with and without earlier surge discharges; the capability with prior discharge, which is always lower, should be used. If the TOV duty exceeds the arrester capability, then a higher voltage rating must be chosen for the arrester, accepting lower protective margins. It is difficult to mitigate TOVs as they typically arise from high-level system design choices.

13.4 Selection by Energy Rating Surge arresters have a switching surge energy discharge capability higher than implied by the low- current, long-duration discharge test. The vendor may provide this value as kJ per kV of MCOV, or rated voltage with footnotes that specify the number, shape, and interval between surges. For example, if the

13-8

Power Systems

capability is given as 8 kJ per kV of MCOV for a single discharge, then a 108 kV arrester with an MCOV of 84 kV has a switching surge discharge capability of 672 kJ. The energy capability may be increased by using two or more columns of metal oxide in parallel. The manufacturer has to match the disks to achieve this and typically two parallel columns provide a bit less than double the energy rating. Note that increasing the arrester voltage rating will not help much, because even though the energy rating increases, the discharge voltage also increases resulting in higher surge energy discharge. Parallel columns will not help with TOV, as TOV is a sustained phenomena supported by the system voltage. There is some evidence that these energy ratings are actually the point at which failure probability becomes nonzero. Hileman (1999) estimated the probability of failure from several reported tests, assuming a Weibull function to model the failure probability PF = 1 − 0.5(

( z/4 ) +1)

Z=

5

WC /WR − 2.5 0.375

(13.3)

where WR is the energy rating WC is the calculated energy duty Under this model, P F reaches 50% only at 2.5 times the energy rating. Note that standards and vendors do not address this point yet, but consideration of the actual failure probability may avoid overdesign for switching surge energy discharge. A conservative estimate of the switching surge energy discharge, WC, is

1 WC = CE 2 2

(13.4)

For energy in joules, C is the total capacitance in μF and E is the peak voltage in kV. The peak voltage ranges from 2 to 2.5 pu of the peak line-to-neutral voltage, although it may be higher for cases like capacitor bank restrikes and line reclosing. The total capacitance may come from an overhead line, a cable, or a shunt capacitor bank. Better estimates may be obtained through time-domain simulation. Lightning surge discharge also poses a significant energy duty on surge arresters. However, the switching surge energy rating does not apply directly and the duty might be better described as a charge duty. In either case, the lightning discharge current may be represented as a decaying exponential with total area equal to the stroke charge. Details of the wave front are unimportant for this evaluation. If one arrester discharges all of the lightning stroke current, a conservative estimate of the energy duty would be

WC = QEd

(13.5)

where Q is the total stroke charge in coulombs Ed is the arrester discharge voltage at the peak lightning stroke current The actual energy discharge is reduced by nonlinear arrester characteristics and by sharing from nearby arresters. Even though nearby arresters will not have matched characteristics, they still share a significant portion of the energy because of inductances between the arresters and also because the characteristics are more linear at lightning discharge current levels than at switching discharge current levels. Time-domain simulation helps to quantify these effects.

13-9

Surge Arresters

13.5 Arrester Modeling Surge arresters are both nonlinear and frequency-dependent devices. The only strictly correct model would be a time-domain simulation at the level of material physics. In most practical cases, one must use a nonlinear resistance with a separate linear circuit to represent frequency-dependent effects. Surge arresters also interact with other equipment that has distributed parameters. Usually the other system nonlinearities are not important, but in special cases some of these effects (e.g., corona) may become significant. In general, arrester application studies need to be done with time-domain simulation in the EMTP or one of its variants like the Alternative Transients Program (ATP). In the absence of EMTP or ATP, the arrester discharge voltage may be approximated for a given surge current peak and waveshape. First, choose the FOW, 8 × 20 μs, or switching surge characteristic that best matches the waveshape. If the manufacturer’s data are not available, Figures 13.2 and 13.3 may be used with a normalizing value (10 kA discharge) of 2.3 times MCOV. In cases where the discharge current peak is “known,” interpolate with this value to estimate the peak discharge voltage. More often, though, a voltage surge arrives at the arrester location and the discharge current peak is not known in advance. In the simplest case, the arrester is at the end of a line with known surge impedance, Z, incoming surge magnitude, E, and power frequency offset, Vpf. The arrester current and voltage may then be estimated according to Hileman (1999) IA =

2E − E0 − Vpf Z + RA

Ed = E0 + I A RA

(13.6)

EA = Ed + Vpf

This approximates the arrester discharge characteristic, Ed, with a straight line having slope R A and intercept E0, as determined by interpolation on either the manufacturer’s data or Figures 13.2 and 13.3. The actual arrester voltage, EA, is the discharge voltage, Ed, plus the power frequency offset voltage, Vpf. To use Equation 13.6, make the first guess at IA and determine the corresponding E0 and R A parameters from the discharge characteristic. Then solve for IA using the first equation; if it falls outside the linear segment used to estimate R A and E0, iterate until convergence on IA. Then Ed and EA may be determined directly. Hileman (1999) presents more detailed estimating methods for other situations, but in those cases it may be easier to use EMTP or ATP. Figure 13.5 shows three frequency-dependent surge arrester models. The Cigre model (Hileman et al., 1990) works well, but it requires the user to program a time-dependent conductance to simulate the arrester turn-on characteristics. It is not directly applicable to EMTP or ATP. The IEEE model (IEEE PES task force) uses high-frequency and low-frequency discharge characteristics represented by A0 and A1, respectively, with a two-stage RL filter connecting them. At nominal voltage, the arrester current is primarily capacitive, represented by the shunt C component. Both IEEE and Cigre models require

L

G(t )

L0

L1

R0

R1 A0

C

Cigre

L1

L0

A1

IEEE

FIGURE 13.5 Frequency-dependent surge arrester models.

R0

A0

Simplified IEEE

A1

13-10

Power Systems

iteration of the parameters to represent a particular arrester, before it can be used in application studies. A simplified version of the IEEE model (Magro et al., 2004) has been developed for use “out of the box”; it ignores the capacitive effect. In the Cigre model, the turn-on conductance begins at zero and increases with time. A reasonable approximation for distribution arresters is dG Gref = dt T

2 U G G I + + 1 1 exp G U ref ref Gref I ref

G0 = 0 T = 80 Gref =

(13.7)

34 U10

I ref = 5.4

U ref = kU10

where U10 is the 10 kA discharge voltage in kV U is the arrester voltage in kV I is the arrester current in kA k is a constant ranging between 0.03 and 0.05, depending on the manufacturer The series inductance, L, is approximately 1 μH per meter of arrester height for outdoor arresters, and one-third of that for gas-insulated substation (GIS) arresters. The following are starting values for the IEEE model parameters:

L1 = 15

d n

R1 = 65

d n

L0 = 0.2

d n

R0 = 100

d n

C = 100

n d

where d is the arrester height in meters n is the number of metal oxide columns in parallel L is in μH R is in Ω C is in pF

(13.8)

13-11

Surge Arresters

Parameters for the simplified model are R0 = 1e6

L0 =

1 VFOW − V10 Vn 12 V10

L1 =

1 VFOW − V10 Vn 4 V10

(13.9)

The inductance values L0 and L1 are in μH, VFOW is the FOW discharge voltage in kV, V10 is the 8 × 20 μs discharge voltage in kV, and Vn is the arrester voltage rating (not the MCOV), also in kV. R0 is present for numerical stability and may not be required for some time-domain simulators. R0 may also be increased by one or two orders of magnitude for extra high voltage (EHV) levels. Whenever VFOW is not available or when the ratio VFOW/V10 is more than 1.18, the simplified inductance parameters become L0 = 0.01Vn

L1 = 0.03Vn

(13.10)

In both IEEE models, A1 closely follows the arrester discharge voltage characteristic for 8 × 20 μs discharge currents. A0 is 20%–30% above A1, such that in parallel and considering the RL filter at high frequency, they approximate the arrester’s FOW protective level (FOWPL). At low frequency, the filter inductances have little impact and A1 dominates the model characteristic. All of the frequency-dependent models include built-in arrester inductance. The arrester lead and pedestal should also be modeled, along with bus, line, or cable lengths to the nearby protected equipment. The lead and pedestal may be represented either as lumped inductance, approximately 1 μH per meter, or with traveling wave lines having surge impedance of about 300 Ω. The separation effect may be approximated for an open-end termination, at which traveling wave reflections will cause doubling before the arrester’s limiting effect is seen. The peak voltage at the open end is

E T = E d + 2S τ

(13.11)

where Ed is the arrester discharge voltage in kV S is the surge steepness in kV/μs τ is the travel time in μs to the open line end For overhead lines or buswork, τ is approximately the distance in meters, divided by 300. For example, if the arrester discharge voltage is 297 kV, the surge steepness is 1000 kV/μs and the distance to the protected transformer is 10 m, then the peak transformer voltage, ET, is approximately 364 kV. That is an increase of 22% from 297 kV, illustrating the importance of minimizing lead lengths. Note that this estimate ignores the effect of transformer capacitance and power frequency offset, both of which would further increase ET. In many cases, the ground resistance does not have to be included. The main exceptions occur when arresters are applied outside of the substation, such as riser poles, distribution transformers, and transmission line towers. In the riser pole application, a ground connection affects the differential-mode surge entering the cable. On distribution lines, the multigrounded neutral conductor allows current to circulate back through nearby arresters. The same thing happens with transmission line arresters installed with overhead shield wires. In all these cases, the ground resistances should be modeled, along with any conductors that are grounded periodically.

13-12

Power Systems

If it is necessary to model an older SiC arrester, the 8 × 20 μs discharge characteristic is represented very well by a single exponential power law

I d = kVdα

(13.12)

This must be connected in series with a gap, having a sparkover voltage ranging anywhere from 5% to 20% higher than V10. The exponential power, α, ranges from two to six, with higher values being more typical for SiC surge arresters. If the actual data are no longer available, it may be assumed that the 10 kA discharge voltage is approximately equal to the 10 kA, 8 × 20 μs discharge voltage of a metal oxide arrester having the same voltage rating (Sakshaug, 1991). On that assumption, the value of k becomes, for units of kA and kV,

k=

10 V10α

(13.13)

Typical SiC data may also be found in Greenwood (1991). Both the FOW and switching surge protective levels for SiC arresters would be significantly higher than for metal oxide arresters of the same voltage rating, even when the V10 values are approximately equal. Conversely, the α parameter in metal oxide varies with current, so that a single exponential does not provide a good model. Piecewise exponential segments are often used for metal oxide surge arrester models in EMTP and ATP.

13.6 Applications Perhaps the most important surge arrester application is to protect the equipment in a substation. After selecting the lowest possible arrester voltage rating, one typically applies an arrester as close as possible to each transformer terminal. Then given the transformer BIL, a protective margin is calculated using the 8 × 20 μs discharge voltage at the “coordination current,” which is typically 10 kA:

BIL − 1 M pct = 100 V10

(13.14)

where Mpct is the protective margin, in percent BIL is the transformer basic insulation level, in kV V10 is the arrester’s 10 kA discharge voltage, in kV The margin could also be calculated at a different coordination current, such as 5 or 20 kA, in which case V5 or V20 would be used instead of V10. One could also calculate a chopped wave protective margin, using the chopped wave insulation level and FOWPL in place of BIL and V10. Finally, one could calculate a margin for switching surges using the BSL and the arrester’s switching surge protective level. In all cases the principle applied is the same and 10%–20% should be considered the minimum acceptable margin. These margins are usually easy to obtain. For example, on a 138 kV system, the transformer BIL is at least 450 kV. If the system is effectively grounded, a 108 kV arrester might be used with a 10 kA discharge voltage of 263 kV, providing a margin of 71%. Using a 120 kV arrester with 297 kV discharge voltage, the margin is still 52%. As discussed in Chapter 14, many uncertainties and deviations complicate the use of margins. After a long time in service, the insulation strength may not really be at the original BIL and BSL levels. The actual surges also differ from laboratory test waves, so that the BIL and BSL do not describe the insulation strength perfectly. Furthermore, several important electrical circuit phenomena are not considered in the simple use of margins and Figure 13.6 illustrates some of them.

13-13

Surge Arresters

S (kV/μs)

E

LT

LB –Vpf

Breaker

Lead + pedestal

Transformer

Z/(n – 1)

FIGURE 13.6 Station protection by surge arresters.

The surge arrester’s lead and pedestal connections should be added to the basic arrester model, either as lumped inductances or as very short distributed-parameter lines. As a result, the voltage to ground at the point of arrester connection will be somewhat higher than the pure arrester discharge voltage. The distances to protected equipment, LB and LT in Figure 13.6, should also be represented with distributed-parameter lines. The arrester itself should be represented with a frequency-dependent model, as discussed earlier. At these high frequencies, the transformer should be represented as a capacitance ranging from 1 to 10 nF, with some sensitivity analysis of this parameter. The incoming surge itself is primarily described with a steepness, S, ranging from 1000 to 2000 kV/μs, depending on the transmission line characteristics, lightning environment, and design level of mean time between failures. The surge magnitude, E, is typically based on some percentage of the line insulation critical flashover (CFO) voltage. It arrives on top of an opposite-polarity power frequency offset, so the arriving surge voltage to ground is actually E − Vpf. All of these factors tend to increase the peak voltage appearing at the transformer, which decreases the protective margin. Note: when using EMTP, the margin is calculated from the simulated peak voltage at the transformer, instead of V10. If two or more lines enter the station, which is typically the case, the additional lines tend to increase protective margins. As an approximation, these lines may be represented with the line surge impedances to ground, connected in parallel at the point of arrester connection. This is shown in Figure 13.6, where n is the total number of lines entering the station. A better representation would be to represent the bus sections with distributed-parameter lines, and connect the line entrances at points where they actually enter the station. A further refinement is to represent the surge with a Thevenin equivalent. The simplest such equivalent is a surge voltage of 2E on top of a DC voltage equal to −Vpf, both behind a lumped resistance equal to the line surge impedance, Z. The next refinement would be to represent the line with distributed parameters back to the stroke point, with a tower footing resistance value in place of Z. Even with these refinements, a single-phase model is adequate for the simulation. Sometimes, the switching surge energy discharge duty on a station arrester will exceed its rating. Typically this happens when large shunt capacitor banks, or long cables, are nearby. The arrester can be ordered with parallel columns, which are matched at the time of manufacturing to achieve effective sharing of the energy duty. If not matched, a small difference in the discharge characteristic could result in one column discharging virtually all of the energy. This would happen even with typical small manufacturing variations of 3% or less. Once the model in Figure 13.6 has been set up, the arrester rating and location will be finalized to adequately protect each transformer. Protective margins for other equipment will also be available, such as the circuit breaker in Figure 13.6. For many stations, arresters at the transformer terminals will adequately protect all equipment in the station. If not, arresters can be added to protect equipment that is far away from the transformers. This process applies to each nominal voltage level in the substation. As a further constraint, the arresters on two different voltage levels should be coordinated so that the high-side arrester always operates first for a surge impinging on the high side of the transformer. If the

13-14

Power Systems

FIGURE 13.7 Transmission line protection by surge arresters.

low-side arresters operate first, they might discharge all of the energy in the high-side surge, leading to possible failure. This coordination is achieved by ensuring 4% margin between the transformer turns ratio and the ratio of arrester protective levels. For example, a 345/138 transformer has a turns ratio of 2.5. The switching surge discharge voltage of the high-side arrester should be no more than 2.5/1.04, or 2.4 times the switching surge discharge voltage of the low-side arrester. Figure 13.7 shows how surge arresters may be used to protect overhead transmission lines. Arresters on the line side of a circuit breaker, shown to the left in Figure 13.7, can protect the breaker from lightning surges traveling in from the line, while the breaker is open. If the breaker is open for maintenance, a disconnect switch would be opened to isolate it from the line. The need for protection would arise when the breaker opens to clear a lightning-induced fault during a storm, and then a subsequent lightning stroke to the line causes another surge while the breaker is still open. As this condition occurs rarely, some utilities may not use arresters to provide this protection. On long EHV transmission lines, surge arresters may be applied at each end to mitigate switching surges during line energization and reclosing. A statistical switching study using either EMTP or a transient network analyzer (TNA) is required for this application. The overvoltages will be higher at intermediate points along the line, due to separation effects. The study results should include a profile of probabilistic switching overvoltage parameters vs. line length, to be used with a probabilistic evaluation of the switching surge flashover rate. Commonly used alternatives to these line-end arresters include preinsertion resistors in the circuit breakers or controlled-closing circuit breakers. While often more expensive and more complicated, these alternatives typically produce a flatter overvoltage profile along the line length. Some EHV transmission lines use series capacitors to reduce their electrical length, and during faults on the line, very high voltages would result on the capacitors. Metal oxide surge arresters typically protect these capacitor banks, as shown in Figure 13.7. The energy dissipation requirements are severe and the manufacturer takes extra care to ensure that the parallel metal oxide columns are matched so that each column shares the energy and current duty. The blocks may also be larger than in typical stationclass surge arresters. Although the transient phenomena is at power frequency and subsynchronous frequencies, these energy calculations are normally done using EMTP. Recently, arresters have been applied to transmission line towers to provide lightning protection of the line insulation. These may be used in place of, or in addition to, overhead shield wires. Some manufacturers have special-purpose line arresters, which are comparable to intermediate or distribution heavy-duty class. Arresters may be used in addition to a shield wire in areas where low footing resistance is difficult to achieve or in high-exposure areas like river crossings. In that case, the arrester energy duty is lessened because the shield wire and footing discharge most of the lightning stroke current. Without a shield wire, the energy duties will increase but are still mitigated by sharing from the arresters on nearby towers. Arresters on transmission towers can have high protective levels, well above the anticipated switching surge and TOV levels, but still low enough to protect the line insulation during lightning stroke discharges. As a result, these line arresters have a smaller block diameter and lower energy discharge capability than station class arresters. The protective level of an EGLA is approximately equal to the gap sparkover voltage, or the 1 kA discharge voltage of the metal oxide. However, success or failure of this application depends more on the ability to withstand vibration and other mechanical stresses, rather than electrical performance.

13-15

Surge Arresters Isurge

Lead House load Service drop

Pole ground

Transformer

House ground

FIGURE 13.8 Distribution transformer protection by surge arresters.

Surge arresters protect virtually all distribution transformers connected to overhead lines, as illustrated in Figure 13.8. It is essential to keep the lead length short. The historical rules of thumb like “2 kV per foot” are grossly understated, because they were based on laboratory test waveshapes that do not represent natural lightning surges. If the arrester is installed between a small fuse and the transformer, then the arrester discharge current must also flow through the fuse, which may cause the fuse to melt unnecessarily. If this becomes a problem, larger fuses or completely self-protected transformers (CSPs), which have internal fuses, could help mitigate it. Because the arrester is connected between the transformer tank and terminal, pole ground resistance and pole download inductance do not have a direct impact on the transformer’s protective margin. However, if the pole ground resistance is too high with respect to the house or customer ground resistance, higher surge currents will circulate in the secondary service drop and the customer load equipment. This may lead to failures due to surges in the transformer secondary, or damage to customer equipment, even when the transformer primary enjoys a comfortable protective margin. It has been suggested that transformers with interlaced secondary windings are less susceptible to these failures. In any case, the secondary surges should be limited by using triplex service drops rather than open-wire drops, and also by keeping the utility pole ground resistance to a reasonably low value. Given the dispersed installation of transformers on a distribution feeder, surge arresters will be installed at relatively close spacings on the distribution line. Typical average spacings might be from two to four poles in built-up areas, although not necessarily on all phases. These arresters may be used to provide some protection from lightning flashovers of the line insulation. Because of the low insulation levels, typically 100–300 kV CFO, this protection will generally be more effective for induced voltages from nearby lightning strokes, which do not actually hit the line. For direct strokes to the line, energy duty on the arresters may be too high, although there is little field experience to indicate that arresters are failing at high rates due to lightning. If the arresters are not placed at every pole and on every phase, separation effects will be more significant than in a typical substation. Some trials have been done with arresters on just the topmost phase, reasoning that the arrester will convert that phase conductor into a shield wire during a lightning discharge. However, the topmost phase is typically not high enough to serve as an effective shield wire for the two outside phase conductors. Many distribution engineers feel that a direct stroke to the line will result in either a line insulation flashover, or less often, an arrester failure. Figure 13.9 shows the application of surge arresters at the riser pole, where an underground cable segment connects to an overhead primary distribution feeder. The riser pole arrester limits a surge entering the cable, but surge doubling may occur at the cable end and tap points. If the lightning surge is “bipolar,” the surge may even quadruple at the open cable ends (Barker, 1990). If the nominal voltage and cable lengths are low enough and setting aside the small risk of voltage quadrupling, the riser pole arrester may be sufficient. For longer cables, it may be necessary to add arresters at some of the open cable ends

13-16

Power Systems

Riser

Scout

Scout

Tap End End

FIGURE 13.9 Riser pole and scout arresters on a distribution system.

or tap points. Before taking that step, the effect of “scout arresters,” shown at either end of the overhead feeder section in Figure 13.9, should be included in the study. These arresters will mitigate most surges arriving at the riser pole.

References Barker, P.P., Voltage quadrupling on a UD cable, IEEE Trans. Power Delivery, 5(1), 498–501, January 1990. Greenwood, A.N., Electrical Transients in Power Systems, chap. 16, Protection of systems and equipment against transient overvoltages. 2nd edn., John Wiley & Sons, New York, 1991, pp. 533–534. Hamel, A. and St. Jean, G., Comparison of ANSI, IEC, and CSA standards’ durability requirements on station-type metal oxide surge arresters for EHV power systems, IEEE Trans. Power Delivery, 7(3), 1283–1298, July 1992. Hileman, A.R., Insulation Coordination for Power Systems, Marcel Dekker, Inc., New York, 1999, pp. 497–675. Hileman, A.R., Roguin, J., and Weck, K.H., Metal oxide surge arresters in AC systems—Part v: Protection performance of metal oxide surge arresters, Electra, 133, 133–144, December 1990. IEC Standard 60099-1, Edition 3.1, Surge Arresters—Part 1: Nonlinear Resistor Type Gapped Surge Arresters for A.C. Systems, International Electrotechnical Commission, December 1999. IEC Standard 60099-3, Edition 1.0, Surge Arresters—Part 3: Artificial Pollution Testing of Surge Arresters, International Electrotechnical Commission, September 1990. IEC Standard 60099-4, Edition 2.2, Surge Arresters—Part 4: Metal Oxide Surge Arresters without Gaps for AC Systems, International Electrotechnical Commission, May 2009. IEC Standard 60099-5, Edition 1.1, Surge Arresters—Part 5: Selection and Application Recommendations, International Electrotechnical Commission, March 2000. IEC Standard 60099-6, Edition 1.0, Surge Arresters—Part 6: Surge Arresters Containing Both Series and Parallel Gapped Structures—Rated 52 kV and Less, International Electrotechnical Commission, August 2002. IEC Standard 60099-7, Edition 1.0, Surge Arresters—Part 7: Glossary of Terms and Definitions, International Electrotechnical Commission, April 2004. IEC Standard 60099-8, Edition 1.0, Surge Arresters—Part 8: Metal-Oxide Surge Arresters with External Series Gap (EGLA) for Overhead Transmission and Distribution Lines of A.C. Systems above 1 kV, International Electrotechnical Commission, January 2011. IEC Standard 61643-1, Edition 2.0, Low-Voltage Surge Protective Devices—Part 1: Surge Protective Devices Connected to Low-Voltage Power Distribution Systems—Requirements and Tests, International Electrotechnical Commission, May 2003. IEC Standard 61643-12, Edition 2.0, Low-Voltage Surge Protective Devices—Part 12: Surge Protective Devices Connected to Low-Voltage Power Distribution Systems—Selection and Application Principles, International Electrotechnical Commission, November 2008. IEEE PES Task Force on Data for Modeling System Transients, Parameter determination for modeling system transients—Part v: Surge arresters, IEEE Trans. Power Delivery, 20(3), 2073–2078, July 2005.

Surge Arresters

13-17

IEEE Std. C62.11-2005, IEEE Standard for Metal-Oxide Surge Arresters for AC Power Circuits (>1 kV), Institute of Electrical and Electronic Engineers, March 2006. IEEE Std. C62.11a-2008, IEEE Standard for Metal-Oxide Surge Arresters for AC Power Circuits (>1 kV) Amendment 1: Short-Circuit Tests for Station Intermediate and Distribution Arresters, Institute of Electrical and Electronic Engineers, July 2008. IEEE Std. C62.22-2009, IEEE Guide for the Application of Metal-Oxide Surge Arresters for AlternatingCurrent Systems, Institute of Electrical and Electronic Engineers, July 2009. IEEE Std. C62.41.1-2002, IEEE Guide on the Surge Environment in Low-Voltage (1000 V and Less) AC Power Circuits, Institute of Electrical and Electronic Engineers, November 2002. IEEE Std. C62.41.2-2002, IEEE Recommended Practice on Characterization of Surges in Low-Voltage (1000 V and Less) AC Power Circuits, Institute of Electrical and Electronic Engineers, November 2002. IEEE Std. C62.72-2007, IEEE Guide for the Application of Surge-Protective Devices for Low-Voltage (1000 V or Less) AC Power Circuits, Institute of Electrical and Electronic Engineers, August 2007. IEEE Std. C62.1-1989, IEEE Standard for Gapped Silicon Carbide Surge Arresters for AC Power Circuits, Institute of Electrical and Electronic Engineers, April 1990. IEEE Std. C62.2-1987, IEEE Guide for the Application of Gapped Silicon-Carbide Surge Arresters for Alternating-Current Systems, Institute of Electrical and Electronic Engineers, April 1989. Magro, M.C., Giannettoni, M., and Pinceti, P., Validation of ZnO surge arrester model for overvoltage studies, IEEE Trans. Power Delivery, 19(4), 1692–1695, October 2004. Osterhout, J.C., Comparison of IEC and U.S. standards for metal oxide surge arresters, IEEE Trans. Power Delivery, 7(4), 2002–2011, October 1992. Sakshaug, E.C., A brief history of AC surge arresters, IEEE Power Eng. Rev., 40, 11–13, August 1991. Sakshaug, E.C., Kresge, J.S., and Miske, S.A., A new concept in arrester design, IEEE Trans. Power Appar. Syst., 96(2), 647–656, March/April 1977.

14 Insulation Coordination 14.1 Insulation Coordination................................................................. 14-1 14.2 Insulation Characteristics.............................................................. 14-2 14.3 Probability of Flashover.................................................................. 14-2 Multiple Gaps per Phase • Multiple Gaps and Multiple Phases

14.4 Flashover Characteristics of Air Insulation................................ 14-3

Stephen R. Lambert Shawnee Power Consulting, LLC

Voltage Waveshape • Electrode Configuration • Effect of Insulator • Effect of Atmospheric Conditions on Air Insulation • Altitude • Insulator Contamination • Application of Surge Arresters • Examples of Surge Arrester Application (Nonself-Restoring Insulation)

References................................................................................................... 14-11

14.1 Insulation Coordination The art of correlating equipment electrical insulation strengths with expected overvoltage stresses so as to result in an acceptable risk of failure while considering economics and operating criteria (McNutt and Lambert, 1992). Insulation properties can be characterized as self-restoring and nonself-restoring. Self-restoring insulation has the ability to “heal” itself following a flashover, and such insulation media is usually associated with a gas—air, SF6, etc. Examples include overhead line insulators, station buswork, external bushing surfaces, SF6 buswork, and even switchgear insulation. With self-restoring insulation, some flashovers are often acceptable while in operation. An EHV transmission line, for example, is allowed to experience occasional line insulator flashovers during switching operations such as energizing or reclosing, or as a result of a lightning flash striking the tower, shield wires, or phase conductors. Nonself-restoring insulation is assumed to have permanently failed following a flashover, and repairs must be effected before the equipment can be put back into service. Insulation such as oil, oil/paper, and solid dielectrics such as pressboard, cross-link polyethylene, butyl rubbers, etc., are included in this insulation class. Any flashover of nonself-restoring insulation, say within a transformer or a cable, is unacceptable as such events usually result in lengthy outages and costly repairs. The performance level of self-restoring insulation is usually addressed and defined in terms of the probability of a flashover. Thus, for a specific voltage stress, a given piece of insulation has an expected probability of flashover (pfo), e.g., a 1-m conductor-to-conductor gap exposed to a 490-kV switching surge would be expected to have a 50% chance of flashover; with a 453-kV surge, the gap would be expected to have a 10% chance of flashover, etc. Consequently, when self-restoring insulation is applied, the procedure is to select a gap length that will give the overall desired performance (pfo) as a function of the stress (overvoltages) being applied.

14-1

14-2

Power Systems

For nonself-restoring insulation, however, any flashover is undesirable and unacceptable, and consequently for application of nonself-restoring insulation, a capability is selected such that the “100%” withstand level (effectively a 0% chance of flashover) of the insulation exceeds the highest expected stress by a suitable margin.

14.2 Insulation Characteristics Self-restoring (as well as nonself-restoring) insulation has, when exposed to a voltage, a pfo which is dependent on • • • • •

Dielectric material (air, SF6, oil…) Waveshape of the stress (voltage) Electrode or gap configuration (rod–rod, conductor to structure…) Gap spacing Atmospheric conditions (for gases)

14.3 Probability of Flashover Assuming the flashover characteristics of insulation follow a Gaussian distribution, and this is a good assumption for most insulation media (air, SF6, oil, oil/paper), the statistical flashover characteristics of insulation can be described by the V50 or mean value of flashover, and a standard deviation. The V50 is a function of the rise time of the applied voltage, and when at a minimum, it is usually known as the CFO or critical flashover voltage. Consequently, for a given surge level and insulation characteristic, the pfo of a single gap can be described by p, and can be determined by first calculating the number of standard deviations the stress level is above or below the mean: #δ =

Vstress − V50 1 standard deviation

(14.1)

For air insulation, 1 standard deviation is either 3% of the V50 for fundamental frequency (50–60 Hz) voltages and for lightning impulses or 6% of the V50 for switching surge impulses. That the standard deviation is a fixed percentage of the V50 and is not a function of gap length is very fortuitous and simplifies the calculations. Once the number of standard deviations away from the mean has been found, then by calculation or by entering a table, the probability of occurrence associated with that number of standard deviations is found. Example: Assume that an insulator has a V50 of 1100 kV with a standard deviation of 6%, and a switching overvoltage of 980 kV is applied to the insulation. The stress is 1.82 standard deviations below the mean: #δ =

980 − 1100 0.06 × 1100

= −1.82 standard deviations below the mean

(14.2)

By calculation or table, the probability associated with −1.82 standard deviations below the mean (for a normal distribution) is 3.4%. Thus, there is a 3.4% chance of insulation flashover every time the insulation is exposed to a 980-kV surge. The physics of the flashover mechanism precludes a breakdown or flashover below some stress level, and this is generally assumed to occur at 3.5–4 standard deviations below the mean.

14-3

Insulation Coordination

14.3.1 Multiple Gaps per Phase The pfo, Pn, for n gaps in parallel (assuming the gaps have the same characteristics and are exposed to the same voltage) can be described by the following equation where p is the pfo of one gap. This mathematical expression defines the probability of one or more gaps flashing over, but practically only one gap of the group will flashover as the first gap to flashover reduces the voltage stress on the other gaps: Pn = 1 − (1 − p)n

(14.3)

14.3.2 Multiple Gaps and Multiple Phases Analysis of some applications may not only require consideration of multiple gaps in a given phase but also of multiple phases. Consider the pfo analysis of a transmission line; during a switching operation for example, multiple towers are exposed to surges and at each tower, each of the three phases is stressed (typically by different surge magnitudes). Thus it is important to consider not only the multiple gaps associated with the multiple towers, but also all three phases often need to be considered to determine the overall line pfo. The overall pfo for a given surge, PFO, can be expressed as

(

PFO = 1 − 1 − pfon,a

) (1 − pfo ) (1 − pfo ) g

g

n,b

n,c

g

(14.4)

where pfon,x is the pfo of the x phase for the given surge n, g is the number of towers (gaps in parallel) The simultaneous analysis of all three phases can be important especially when various techniques are used to substantially suppress the surges (Lambert, 1988).

14.4 Flashover Characteristics of Air Insulation 14.4.1 Voltage Waveshape Waveshapes used for testing and for determining the flashover response of insulation have been standardized by various groups and while there is not 100% agreement, the waveshapes used generally conform to the following: Fundamental frequency Switching impulse Lightning impulse

50 or 60 Hz sine wave (8000 μs rise time) 200–250 μs by 2000 μs 1.2 μs by 50 μs

The impulse waveshapes are usually formed by a double exponential having the time to crest indicated by the first number and the time to 50% of the crest on the tail of the wave indicated by the second number. Thus, a lightning impulse would crest at 1.2 μs and following the crest would fall off to 50% of the crest at 50 μs. Fundamental frequency characteristics have been published, and typical values are indicated on Figure 14.1 (Aleksandrov et al., 1962; EPRI, 1982). Equations have also been published or can be developed which define the typical responses to positive polarity switching and lightning impulses (see Figure 14.2). Insulation usually has a lower withstand capability when exposed to positive polarity impulses than when exposed to negative impulses; thus, designs are usually based on positive magnitude impulses.

14-4

Power Systems

Critical flashover voltage (kVcr)

2500 Conductor to conductor

2000 1500

Insulator string

1000

Rod to plane

500 0

0

1

2

3

4 5 Gap spacing (m)

6

7

8

9

FIGURE 14.1 V50 for fundamental frequency waveshapes. (From Transmission Line Reference Book, 345 kV and Above, 2nd edn., Electric Power Research Institute, Palo Alto, CA, 1975.) 6000 5000

Lightning impulses

V50 (kV)

4000

Switching impulses

3000 2000 1000 0

0

2

4

6 Gap spacing (m)

8

10

12

FIGURE 14.2 V50 for impulses—positive polarity, rod-plane gap. (From EHV Transmission Line Reference Book, Edison Electric Institute, New York, 1968; Gallet, G. et al., IEEE Trans. Power Appar. Syst., PAS-95, 580, 1976.)

For switching surge impulses (gaps ≤ 15 m) (Gallet et al., 1976):

V50 = k

3400 kV 1 + (8/d )

(14.5)

For lightning impulses, the following equation can be developed from EEI (1968):

V50 = k * 500 * d kV

where k is an electrode factor reflecting the shape of the electrodes (Paris, 1967) d is the electrode gap spacing in meters

(14.6)

14-5

Insulation Coordination

14.4.2 Electrode Configuration Electrode configuration has a pronounced effect on the V50 characteristics, and this is reflected as a gap or electrode factor, k (Paris, 1967). Examples of k are Rod–plane Conductor–structure Rod–rod Conductor–rope Conductor–rod

1.00 1.30 1.30 1.40 1.65

14.4.3 Effect of Insulator The presence of an insulator in a gap tends to reduce the gap factor from those given above, mainly due to the terminal electrode configuration (and intermediate flanges for multiunit column bus support insulators). The reduction increases with increased gap factor and typical correction values may be found on Figure 14.3. Note that these corrections are subject to variations (Thione, 1984). Rain has little effect on a gap without an insulator; however, rain does reduce the gap factor when an insulator is present. Reductions as high as 20% have been noted; but, in general a reduction of 4%–5% is typical (Thione, 1984).

14.4.4 Effect of Atmospheric Conditions on Air Insulation V50 for gases is affected by temperature, atmospheric pressure, and humidity, and for air the correction can be expressed as n

δ V50,ambient = V50,NTP Ho

(14.7)

where NTP is the normal temperature and pressure (20°C, 101.3 kPa) Ho is the humidity correction factor n is a gap length correction factor δ is the relative air density correction factor 1

Correction

0.97 String and post insulators

0.94

Post insulators

0.91 0.88 0.85

1

1.1

1.2

1.3 1.4 Gap factor

1.5

1.6

1.7

FIGURE 14.3 Gap factor correction for presence of insulator. (From Thione, L., ELECTRA, 94, 77, 1984.)

14-6

Power Systems

The correction for temperature and pressure, “δ,” is known as the RAD (relative air density) correction factor and is expressed by δ=

0.386H mm of Hg 273 + T

(14.8)

where Hmm of Hg is the atmospheric pressure in mm of Hg T is the temperature in °C The humidity correction factor, Ho, is given in IEEE 4 (1978) and can be expressed approximately by H o ≅ 1.1 − 0.00820 *H AB ≅ 1.1 − 0.008071*VP

(14.9)

where HAB is the absolute humidity in g/m3 VP is the vapor pressure in mm of Hg For switching impulses (and fundamental frequency) the effect of the RAD and humidity on V50 is, however, a function of the gap length and has less effect on longer gap than on shorter gap lengths. For lengths of 0–1 m the n correction factor is 1.0; from 1 to 6 m, the correction decreases linearly from 1.0 to 0.4; and for lengths greater than 6 m, the factor is 0.4. There is no gap length correction for positive lightning impulses (EEI, 1968; EPRI, 1975, 1982). Other approaches for humidity corrections can be found in Menemenlis et al. (1988), Thione (1984), Feser and Pigini (1987).

14.4.5 Altitude Corrections for altitude are also important as the insulation capability drops off about 10% per 1000 m as shown in Figure 14.4. There are various equations for the altitude correction factor (ACF) and the following expression is representative of most in use (IEEE 1312, 1999): ACF = ( e −H t /8600 ) n

(14.10)

where Ht is the altitude in meters n is a gap length correction factor

14.4.6 Insulator Contamination Insulator contamination is an important issue for fundamental frequency voltage considerations, and the equivalent salt density, ESDD, approach is extensively used as a design tool. The contamination severity is defined by the ESDD in mg/cm2, and an insulator creepage distance, in terms of mm/kVrms, phase to phase, can then be selected (IEC 815, 1986). Note that insulator/bushing shed/skirt design has a significant impact on the performance, and some past designs performed poorly due to skirt configuration even though they had large creepage distances. With the ESDD approach insulators are tested to define their expected performance. Table 14.1 shows the relationship between contamination level, ESDD, and recommended creepage distances.

14-7

Insulation Coordination

Altitude correction factor

1.1 Best estimate for contamination

1.0 0.9 0.8

Impulse and fundamental frequency

0.7 0.6

0

500

1000

1500

2000 2500 Altitude (m)

3000

3500

4000

4500

FIGURE 14.4 Altitude correction factors. (From Mizuno, Y. et al., IEEE Trans. Dielectr. Electr. Insul., 4, 286, 1997; IEEE Standard for Insulation Coordination—Part 2, Application Guide, Institute of Electrical and Electronic Engineers (IEEE), 1312, 1999.) TABLE 14.1 Recommended Creepage Distances Contamination Level Light Medium Heavy Very heavy

Minimum Recommended Creepage Distance (mm/ kVrms, phase to phase)

Example

ESDD (mg/cm2)

Low industrial activity Industrial activity—some exposure to wind from the sea Industrial area and areas close to the sea Heavy industrial or sea coast area

0.03–0.06 0.1–0.2

16 20

0.3–0.6 >0.6

25 31

Altitude also has an effect on the performance of contaminated insulation, and the degradation of capability as a function of altitude may be found in Figure 14.4 (Mizuno et al., 1997). Example 14.1 Assume 10 identical substation bus support insulators in parallel located in a 500-kV substation located at sea level; this configuration can be described as an air gap, conductor to rod configuration at standard atmospheric conditions. Assume that an overall pfo for the 10 gaps of 0.5% is desired when the configuration is exposed to a switching surge of 939 kV (2.3 pu on a 500-kV system). What is the required gap clearance in meters? Solution The desired pfo of one gap, p, then should be

pfo10gaps = 0.005 = 1 − (1 − p)10

and p = 1 − (1 − 0.005)1/10

= 0.0005011

14-8

Power Systems From tables or calculations for a normal or Gaussian distribution, a probability of 0.0005011 corresponds to 3.29 standard deviations below the mean (V50). Therefore, the desired V50 value is

939 kV = V50 (1 − 3.29 * 0.06)

and

V50 = 1170 kV

A standard deviation of 6% is often used for the air gap for switching surge stresses. With the V50 of 1170 kV and noting that a conductor to rod gap has a k factor of 1.65 assumed to be reduced to 1.42 due to the presence of the insulator, the desired gap spacing value can be calculated by

1170 kV = 1.42

3400 1 + (8/d )

and

d = 2.56m

14.4.7 Application of Surge Arresters Surge arresters are used to limit overvoltages and as a result, allow reductions in the clearances required for self-restoring gaps (e.g., transmission line towers) as well as the capability required for nonselfrestoring insulation such as transformer windings. In most applications the proper approach is to determine the minimum arrester rating, which can be applied without resulting in damage to the arrester and then to define the insulation level required so as to result in an acceptable pfo or risk of failure. For a transmission line application, for example, although the arrester reduces higher magnitude surges to lower levels, the line is still allowed to have a finite, albeit low, pfo for a specific switching operation. Thus, the arrester, by limiting the higher magnitude surges, allows smaller conductor to tower clearances. However, when arresters are used to protect a transformer for example, an insulation level, which has a significantly higher capability than the maximum surge allowed by the arrester, is selected. This margin between the arrester protective levels (lightning or switching surge) is a function of various considerations as well as the conservatism of the person applying the arrester/insulation system. Today, for new applications, only metal oxide (ZnO) arresters are being applied. Although there are certainly many of the gapped, silicon carbide type arresters still in service and which still perform effectively, in what follows, only metal oxide arresters will be considered to protect insulation. Successful application requires that the arrester survives the electrical environment in which it is placed, and the following arrester capabilities must be carefully considered: MCOV—maximum fundamental frequency continuous operating voltage applied to the arrester TOV—temporary fundamental frequency overvoltages to which the arrester may be exposed Energy—the energy which must be absorbed by the arrester when limiting switching surges 14.4.7.1 MCOV The highest system voltage, which can be continuously applied to the arrester, needs to be determined and the arrester capability, its MCOV rating, should at least be equal to and should usually exceed the highest continuous system voltage by some small margin. For example, if a nominal 345-kV system is

Insulation Coordination

14-9

never operated above 352 kV, then the maximum continuous voltage, which would be expected to be applied to a line to ground arrester, would be 352 kV/ 3 = 203.2 kV. With today’s typical arresters, the next highest available MCOV capability would be 209 kV and is associated with an arrester rated 258 kV. 14.4.7.2 TOV On occasion, the fundamental-frequency voltage applied to an arrester will exceed the expected MCOV. Examples include fault conditions during which line to ground voltages on unfaulted phases can rise significantly (as high as phase-to-phase voltage for ungrounded systems); rise in line voltage when energizing a transmission line (Ferranti effect) and voltages which occur during load rejection events—these are usually associated with voltages experienced on a radial transmission line emanating from a generating plant when the load terminal of the line opens unexpectedly. 14.4.7.3 Energy When an arrester limits switching surges on a transmission line, it can absorb a significant amount of energy, and it can be important to examine events and determine the energy which could be absorbed. Exceeding the arrester’s capability could result in immediate damage to the arrester and failure. It is also important to the arrester’s TOV capability as absorbing energy heats the arrester material, and application of a significant temporary overvoltage immediately following absorption of a significant amount of energy could result in thermal runaway and arrester failure. Following selection of an arrester which would be expected to survive the electrical environment (i.e., the minimum rated arrester), the protective levels of the arrester must be correlated with the insulation capability and acceptable margins between the protective levels and the insulation capability achieved. The protective level or discharge voltage of an arrester is the voltage magnitude to which the arrester will limit the voltage while discharging a surge, and these levels are a function of the waveshape and rise time of the surge as well as the current magnitude of the discharge. In general, the discharge or protective levels considered for coordination with insulation capability are • A 10-kA, 8 × 20-μs discharge for coordination with the insulation full wave or lightning impulse (BIL) capability and • A 0.5–2.0-kA, 36 × 90-μs discharge for coordination with the switching impulse capability There should always be margin between the protective level of the arrester and the insulation capability to allow for uncertainties in arrester protective levels due to surge rise times, discharge currents, and arrester separation distance (faster rise times, higher currents, and longer separation distance or lead lengths generate higher protective levels). Uncertainties in insulation capability include reduced insulation strength due to aging (especially for paper insulation in transformers for example) and limitations of the ability of laboratory dielectric testing to accurately relate to field conditions. In the author’s opinion, a margin of at least 40% is appropriate unless all the uncertainties and the risks are carefully evaluated.

14.4.8 Examples of Surge Arrester Application (Nonself-Restoring Insulation) 14.4.8.1 34.5-kV System Application Surge arresters are to be applied line to ground at the terminals of a circuit breaker (38-kV rating, 150-kV BIL) used on a solidly grounded 34.5-kV system. The highest expected continuous system voltage is 37 kV, and during fault conditions, the phase to ground voltage can rise to 1.4 pu or 27.9 kV. Faults can persist for 20 cycles. The maximum line to ground voltage is 37/ 3 = 21.4 kV and the MCOV of the arrester must meet or exceed this value. An arrester rated 27 kV would be acceptable as it has an MCOV of 22.0 kV. The 1 s TOV

14-10

Power Systems

capability of the arrester is 31.7 kV, and as this exceeds the 27.9 kV phase to ground voltage expected during faults, the 27-kV arrester meets the TOV criteria as well. A 27-kV arrester has a 10-kA discharge level of 67.7 kV, and thus the margin between the discharge or protective level and the insulation BIL is (150/67.7 × 100–100) or 121%. This margin is obviously more than adequate, and selection of an arrester rated 27 kV would be appropriate. 14.4.8.2 500-kV System Application A 500-kV shunt reactor (solidly grounded neutral) is being applied at the end of a 300 km, 500-kV transmission line, and arresters are to be applied line to ground on the terminals of the reactor to limit surges to reasonable levels. The reactor is solidly connected to the line and is switched with the line, and the substation at which the reactor resides is at an altitude of 1800 m. The highest expected continuous system voltage is 550 kV. During line switching operations, the circuit breaker at the reactor terminal may not be closed for some period following energizing of the line/reactor from the other terminal, and the phase to ground voltage at the reactor can be as high as 1.15 pu for as long as 5 min. Arrester energy requirements were determined (by EMTP or TNA simulations of switching operations) to be well within the capability of an arrester rated 396 kV. The minimum required MCOV is 550 / 3 = 317.5kV. The minimum required TOV is 1.15 × 500/ 3 = 332 kV for 300 s, and for most arresters, such a requirement would correlate with a 1 s TOV rating of 451 kV. An arrester rated 396 kV has a 318-kV MCOV and a 1 s TOV rating of 451 kV; thus, a 396-kV arrester would be the minimum rating that could be used. Of course any arrester rated higher than 396 kV could also be used. The 10-kA lightning (8 × 20 μs waveform) and switching surge (2 kA, 36 × 90 μs) discharge levels for a 396-kV and a 420-kV arrester are Discharge Levels Rating (kV) 396 420

10 kA (kV)

Switching Surge (kV)

872 924

758 830

BIL values of 1300 and 1425 kV for the reactor’s internal insulation (i.e., insulation not affected by altitude) could be considered as reasonable candidates for a specification. The corresponding switching impulse levels (SIL) would be 1080 and 1180 kV, respectively, and the following table indicates the margin between the arrester protective levels and the insulation level. 1300-kV BIL Arrester SIL BIL

1425-kV BIL

396 kV (%)

420 kV (%)

396 kV (%)

420 kV (%)

42 49

30 41

56 63

42 54

Application of a 420-kV arrester for a 1300-kV BIL insulation level results in margins below 40%, and unless the application is very carefully considered from the point of view of arrester separation distance and lead length, expected maximum discharge current level, wave rise time, etc., a 396-kV arrester would be a better choice. For a 1425-kV BIL, either the 396-kV or the 420-kV arrester would result in sufficient margins. For external insulation, i.e., the reactor bushings, the effect of altitude on the insulation capability needs to be considered. At 1800 m, the insulation has only 81% of the withstand capability demonstrated at sea level or 0 m. For example, the SIL of a 1425-kV bushing (1180 kV at sea level) would be reduced to 956 kV at 1800 m (1180 × 0.81 = 956 kV), and application of even a 396-kV arrester would result in a margin of 26%—hardly acceptable.

14-11

Insulation Coordination

Assume that a 420-kV arrester was selected to protect the reactor (the arrester itself is rated for application to 3000 m). The switching surge and 10 kA protective levels are 830 and 924 kV, respectively. With a desired minimum margin of 40%, and correcting for altitude, the minimum SIL and BIL at sea level (0 m) should be

Minimum SIL =

830 × 1.4 = 1435 kV 0.81

Minimum BIL =

924 × 1.4 = 1597 kV 0.81

A 1550-kV BIL bushing would have a 1290-kV SIL, and even if one would accept the slightly less than a 36% margin for the BIL, the SIL margin would only be 26%. A 1675-kV BIL bushing would be expected to have a 1390-kV SIL capability, and so the SIL margin would be 36% with a BIL margin of 47%. The next higher rated bushing (1800-kV BIL) would mean applying 800-kV system class bushings, and their increased size and cost would likely not make for a reasonable design. Consequently, specifying a 1675-kV BIL bushing and accepting the slightly reduced SIL margin would be a reasonable compromise. 14.4.8.3 Effect of Surge Reduction Techniques on Overall PFO Application of surge arresters to significantly reduce switching surge levels on transmission line and substation insulators can be effective, however, the designer should be aware that the overall PFO of all three phases needs to be considered as it will usually be higher than that found for a single phase by a factor often approaching three. Also for long transmission lines, application of arresters at the line terminals will certainly limit the surges at the terminals but will not limit the surges at other points on the line to the same level. Consequently, the surge distribution along the line may need to be considered (Lambert, 1988; Ribiero et al., 1991).

References Aleksandrov, G.N., Kizvetter, V.Y., Rudakova, V.M., and Tushnov, A.N., The AC flashover voltages of long air gaps and strings of insulators, Elektrichestvo, 6, 27–32, 1962. EHV Transmission Line Reference Book, Edison Electric Institute, New York, 1968. Feser, K. and Pigini, A., Influence of atmospheric conditions on the dielectric strength of external insulation, ELECTRA, 112, 83–93, 1987. Gallet, G., Bettler, M., and Leroy, G., Switching impulse results obtained on the outdoor testing area at Renardieres, IEEE Transactions on Power Apparatus and Systems, PAS-95(2), 580–585, 1976. Guide for the Selection of Insulators in Respect of Polluted Conditions, The International Electrotechnical Commission Publication 815, 1986. IEEE Standard for Insulation Coordination—Part 2, Application Guide, Institute of Electrical and Electronic Engineers (IEEE) 1312, 1999. IEEE Standard Techniques for High-Voltage Testing, Institute of Electrical and Electronic Engineers (IEEE) 4-1978. Lambert, S.R., Effectiveness of zinc oxide surge arresters on substation equipment probabilities of flashover, IEEE Transactions on Power Delivery, 3(4), 1928–1934, 1988. McNutt, W.J. and Lambert, S.R., Transformer Concepts and Applications Course, Power Technologies, Inc., Schenectady, NY, 1992. Menemenlis, C., Carrara, G., and Lambeth, P.J., Application of insulators to withstand switching surges in substations, part I: Switching impulse insulation strength, 88 WM 077-0, IEEE/PES Winter Meeting, New York, January 31–February 5, 1988.

14-12

Power Systems

Mizuno, Y., Kusada, H., and Naito, K., Effect of climatic conditions on contamination flashover voltage of insulators, IEEE Transactions on Dielectrics and Electrical Insulation, 4(3), 286–289, 1997. Paris, L., Influence of air gap characteristics on line-to-ground switching surge strength, IEEE Transactions on Power Apparatus and Systems, PAS-86(8), 936–947, 1967. Ribeiro, J.R., Lambert, S.R., and Wilson, D.D., Protection of compact transmission lines with metal oxide arresters, CIGRE Leningrad Symposium, 400-6, S33–S91, 1991. Thione, L., Evaluation of the switching impulse strength of external insulation, ELECTRA, 94, 77–95, 1984. Transmission Line Reference Book, 345 kV and Above, 1st edn., Electric Power Research Institute, Palo Alto, CA, 1975. Transmission Line Reference Book, 345 kV and Above, 2nd edn., Electric Power Research Institute, Palo Alto, CA, 1982.

Power System Planning (Reliability)

III

Gerald B. Sheblé

15 Planning Environments Gerald B. Sheblé........................................................................ 15-1

16 Short-Term Load and Price Forecasting with Artificial Neural Networks Alireza Khotanzad................................................................................................................. 16-1

Introduction • Defining a Competitive Framework • Regulated Environment • Other Sections on Planning • References

Artificial Neural Networks • Short-Term Load Forecasting • Short-Term Price Forecasting • References

17 Transmission Plan Evaluation: Assessment of System Reliability N. Dag Reppen and James W. Feltes......................................................................................17-1 Bulk Power System Reliability and Supply Point Reliability • Methods for Assessing Supply Point Reliability • Probabilistic Reliability Assessment Methods • Application Examples • References

18 Power System Planning Hyde M. Merrill......................................................................... 18-1

19 Power System Reliability Richard E. Brown..................................................................... 19-1

Planning Entities • Arenas • The Planning Problem • Planning Processes • References NERC Regions • System Adequacy Assessment • System Security Assessment • Probabilistic Security Assessment • Distribution System Reliability • Typical Sequence of Events after an Overhead Distribution Fault • Distribution Reliability Indices • Storms and Major Events • Component Reliability Data • Utility Reliability Problems • Reliability Economics • Annual Variations in Reliability • References

III-1

III-2

Power System Planning (Reliability)

20 Probabilistic Methods for Planning and Operational Analysis Gerald T. Heydt and Peter W. Sauer.................................................................................... 20-1 Uncertainty in Power System Engineering • Deterministic Power Flow Studies • Monte Carlo Power Flow Studies • Analytical Probabilistic Power Flow Studies • Applications for Available Transfer Capability (ATC) • An Example of Stochastic ATC • An Example of Expected Financial Income from Transmission Tariffs • Wind Energy Resources and Stochastic Power Flow Studies • Solar Photovoltaic Energy Resources and Stochastic Power Flow Studies • Conclusions • References

21 Engineering Principles of Electricity Pricing Lawrence J. Vogt....................................21-1

22 Business Essentials Richard E. Brown............................................................................... 22-1

Electricity Pricing Overview • Electric Cost-of-Service Study • Cost-of-Service Study Framework • Minimum Distribution System Analysis • Analysis of Load Diversity • Analysis of Demand and Energy Losses • Electric Rate Design • Cost Curve Development • Rate Design Methodology • Reference Introduction • Accounting • Finance • Financial Risk • Financial Ratios • Asset Management

Gerald B. Sheblé, executive advisor and senior director of U.S. Research and Development, has more than 38 years of experience in industry, consulting, expert witness, and academic environments. He has led research, development, and implementation for power system operation and planning using classical analytics and computational adaptive agents. Dr. Sheblé is an internationally recognized expert in the area of electric energy auction markets, capital budgeting, power system analysis, and real option valuation. He is nationally recognized as one of the nation’s leading business experts on business trends within the utility industry and on trends forced by external economic drivers. His expertise extends from smart grid, data mining, power system optimization, electric energy market modeling, energy and emission management, control automation, technology innovation, bulk power operations, generation and transmission planning, operator training, energy risk management, distribution operations, energy management systems, to portfolio decision analysis of interchange contracts. His academic work has concentrated on developing concepts, methods, and products for reregulated operation, market analysis and design, equipment scheduling (especially unit commitment), transaction evaluation, reliability analysis by decision analysis, Monte Carlo simulation, failure tree analysis, market simulation and energy economics, decision analysis, and real option valuation of renewable energy equipment. He has pioneered structure and market definition predicting successful industry reregulation. He has also authored over 100 papers and 50 research documents. He is an associate business scene editor of IEEE Power Magazine. His columns are frequently highlighted in respected trade journals for industry innovations. He has been a premier guest or contributor at over 50 specialized workshops or courses throughout Europe and North America in over 24 countries. He has recently been an expert witness on electrical accidents, deficient project management, grounding, and intellectual property rights. He has also been an expert witness on software engineering, software business planning, and industrial trends. Before joining Quanta Technology, he has worked for Commonwealth Edison, Systems Control Inc., Control Data Corporation, Energy and Control Consultants, Auburn University, Iowa State University, Portland Stat University, University of Porto (Portugal), and INESC Porto. Dr. Sheblé received his MBA from The University of Iowa (Henry B. Tippie College of Business), his PhD from Virginia Technological and State University, and his MSEE and BSEE from Purdue University. Dr. Sheblé is an IEEE Fellow for his contributions to the development of auction methods as an alternative to power system optimization methods, addressing the deregulation of the electric utility business. Dr. Gerald Sheblé can be reached at [email protected]

15 Planning Environments 15.1 Introduction..................................................................................... 15-1 15.2 Defining a Competitive Framework............................................. 15-2 Preparing for Competition • Present View of Overall Problem • Economic Evolution • Market Structure • Fully Evolved Marketplace • Computerized Auction Market Structure • Capacity Expansion Problem Definition

Gerald B. Sheblé Quanta Technology, LLC

15.3 Regulated Environment................................................................ 15-12 15.4 Other Sections on Planning......................................................... 15-14 References................................................................................................... 15-15

15.1 Introduction Capacity expansion decisions are made daily by government agencies, private corporations, partnerships, and individuals. Most decisions are small relative to the profit and loss sheet of most companies. (Aggarwal, 1993; Bussey, 1981; Daellenbach, 1984). However, many decisions are sufficiently large to determine the future financial health of the nation, company, partnership, or individual. Capacity expansion of hydroelectric facilities may require the commitment of financial capital exceeding the income of most small countries. Capacity expansion of thermal fossil fuel plants is not as severe, but does require a large number of financial resources including bank loans, bonds for long-term debt, stock issues for more working capital, and even joint-venture agreements with other suppliers or customers to share the cost and the risk of the expansion. This section proposes several mathematical optimization techniques to assist in this planning process. These models and methods are tools for making better decisions based on the uncertainty of future demand, project costs, loan costs, technology change, etc. (Binger, 1998). Although the material presented in this section is only a simple model of the process, it does capture the essence of real capacity expansion problems. This section relies on a definition of electric power industry restructuring presented in Sheblé (1999). The new environment within this work assumes that the vertically integrated utility has been segmented into a horizontally integrated system (Ilic, 1998). Specifically, generation companies (GENCOs), distribution companies (DISTCOs), and transmission companies (TRANSCOs) exist in place of the old. This work does not assume that separate companies have been formed. It is only necessary that comparable services are available for anyone connected to the transmission grid. As can be concluded, this description of a deregulated marketplace is an general version of the commodity markets. It needs polishing and expanding. The change in the electric utility business environment is depicted generically below. The functions shown are the emerging paradigm. This work outlines the market organization for this new paradigm. Attitudes toward restructuring still vary from state to state and from country to country. Many electric utilities in the United States have been reluctant to change the status quo. Electric utilities with high rates are very reluctant to restructure since the customer is expected to leave for the lower prices. 15-1

15-2

Power Systems

Electric utility companies in regions with low prices are more receptive to change since they expect to pick up more customers. In 1998, California became the first state in the United States to adopt a competitive structure, and other states are observing the outcome. Some offer customer selection of supplier. Some offer markets similar to those established in the United Kingdom, Norway, and Sweden, but not Spain. Several countries have gone to the extreme competitive position of treating electricity as a commodity as seen in New Zealand and Australia (Outhred, 1993). As these markets continue to evolve, governments in all areas of the world will continue to form opinions on what market, operational, and planning structures will suit them best.

15.2 Defining a Competitive Framework There are many market frameworks that can be used to introduce competition between electric utilities. Almost every country embracing competitive markets for its electric system has done so in a different manner. The methods described here assume an electric marketplace derived from commodities exchanges like the Chicago Mercantile Exchange (CME), Chicago Board of Trade (CBOT), and New York Mercantile Exchange (NYMEX) where commodities (other than electricity) have been traded for many years. NYMEX added electricity futures to their offerings in 1996, supporting this author’s previous predictions (Sheblé, 1991, 1992, 1993, 1994) regarding the framework of the coming competitive environment. The framework proposed has similarities to the Norwegian-Sweden electric systems. The proposed structure is partially implemented in New Zealand, Australia, and Spain. The framework is being adapted since similar structures are already implemented in other industries. Thus, it would be extremely expensive to ignore the treatment of other industries and commodities. The details of this framework and some of its major differences from the emerging power markets/pools are described in Sheblé (1999). These methods imply that the ultimate competitive electric industry environment is one in which retail consumers have the ability to choose their own electric supplier. Often referred to as retail access, this is quite a contrast to the vertically integrated monopolies of the past. Telemarketers are contacting consumers, asking to speak to the person in charge of making decisions about electric service. Depending on consumer preference and the installed technology, it may be possible to do this on an almost real-time basis as one might use a debit card at the local grocery store or gas station. Real-time pricing, where electricity is priced as it is used, is getting closer to becoming a reality as information technology advances. Presently, however, customers in most regions lack the sophisticated metering equipment necessary to implement retail access at this level. Charging rates that were deemed fair by the government agency, the average monopolistic electric utility of the old environment met all consumer demand while attempting to minimize their costs. During natural or man-made disasters, neighboring utilities cooperated without competitively charging for their assistance. The costs were always passed on to the rate payers. The electric companies in a country or continent were all members of one big happy family. The new companies of the future competitive environment will also be happy to help out in times of disaster, but each offer of assistance will be priced recognizing that the competitor’s loss is gain for everyone else. No longer guaranteed a rate of return, the entities participating in the competitive electric utility industry of tomorrow will be profit driven.

15.2.1 Preparing for Competition Electric energy prices recently rose to more than $7500/MWh in the Midwest (1998) (Midwest ISO, website) due to a combination of high demand and the forced outage of several units. Many midwestern electric utilities bought energy at that high price, and then sold it to consumers for the normal rate. Unless these companies thought they were going to be heavily fined, or lose all customers for a very long time, it may have been more fiscally responsible to terminate services. Under highly competitive scenarios, the successful supplier will recover its incremental costs as well as its fixed costs through the prices it charges. For a short time, producers may sell below their costs, but

Planning Environments

15-3

will need to make up the losses during another time period. Economic theory shows that eventually, under perfect competition, all companies will arrive at a point where their profit is zero. This is the point at which the company can break even, assuming the average cost is greater than the incremental cost. At this ideal point, the best any producer can do in a competitive framework, ignoring fixed costs, is to bid at the incremental cost. Perfect competition is not often found in the real world for many reasons. The prevalent reason is technology change. Fortunately, there are things that the competitive producer can do to increase the odds of surviving and remaining profitable. The operational tools used and decisions made by companies operating in a competitive environment are dependent on the structure and rules of the power system operation. In each of the various market structures, the company goal is to maximize profit. Entities such as commodity exchanges are responsible for ensuring that the industry operates in a secure manner. The rules of operation should be designed by regulators prior to implementation to be complete and “fair.” Fairness in this work is defined to include noncollusion, open market information, open transmission and distribution access, and proper price signals. It could call for maximization of social welfare (i.e., maximize everyone’s happiness) or perhaps maximization of consumer surplus (i.e., make customers happy). Changing regulations are affecting each company’s way of doing business and to remain profitable, new tools are needed to help companies make the transition from the old environment to the competitive world of the future. This work describes and develops methods and tools that are designed for the competitive component of the electric industry. Some of these tools include software to generate bidding strategies, software to incorporate the bidding strategies of other competitors, and updated common tools like economic dispatch and unit commitment to maximize profit.

15.2.2 Present View of Overall Problem This work is motivated by the recent changes in regulatory policies of inter-utility power interchange practices. Economists believe that electric pricing must be regulated by free market forces rather than by public utilities commissions. A major focus of the changing policies is “competition” as a replacement for “regulation” to achieve economic efficiency. A number of changes will be needed as competition replaces regulation. The coordination arrangements presently existing among the different players in the electric market would change operational, planning, and organizational behaviors. Government agencies are entrusted to encourage an open market system to create a competitive environment where generation and supportive services are bought and sold under demand and supply market conditions. The open market system will consist of GENCOs, DISTCOs, TRANSCOs, a central coordinator to provide independent system operation (ISO), and brokers to match buyers and sellers (BROCOs). The long-term planning has been separated under the organization known as regional transmission organization (RTO). The energy market exchange has been absorbed into the ISO instead of being a separate energy mercantile association (EMA). The energy contracts segmented to the distribution companies are often partially segmented to energy service suppliers (ESS), which may include energy service companies (ESCOs) or energy management companies (EMCOs). The services offered to the customers are not different from the traditional average pricing as of this writing. However, the impact of the smart grid is expected to change this dramatically. The interconnection between these groups is shown in Figure 15.1. The ISO is independent and a dissociated agent for market participants. The roles and responsibilities of the ISO in the new marketplace are yet not clear. This work assumes that the ISO is responsible for coordinating the market players (GENCOs, DISTCOs, and TRANSCOs) to provide a reliable power system functions. Under this assumption, the ISO would require a new class of optimization algorithms to perform price-based operation. Efficient tools are needed to verify that the system remains in operation with all contracts in place. This work proposes an energy brokerage model for all services as a novel framework for price-based optimization. The proposed foundation is used to develop analysis and simulation tools to study the implementation aspects of various contracts in a deregulated environment.

15-4

Power Systems

GENCO

BROCO/PX

ISO/EMA/RTO

ESCO

TRANSCO/DISTOCO/ESCO/EMCO/ESS

FIGURE 15.1 New organizational structure.

Although it is conceptually clean to have separate functions for the GENCOs, DISTCOs, TRANSCOs, and the ISO, the overall mode of real-time operation is still evolving. Presently, two possible versions of market operations are debated in the industry. One version is based on the traditional power pool concept (POOLCO). The other is based on transactions and bilateral transactions as presently handled by commodity exchanges in other industries. Both versions are based on the premise of price-based operation and market-driven demand. This work presents analytical tools to compare the two approaches. Especially with the developed auction market simulator, POOLCO, multilateral, and bilateral agreements can be studied. Working toward the goal of economic efficiency, one should not forget that the reliability of the electric services is of the utmost importance to the electric utility industry in North America. In the words of the North American Electric Reliability Council (NERC), reliability in a bulk electric system indicates “the degree to which the performance of the elements of that system results in electricity being delivered to customers within accepted standards and in the amount desired. The degree of reliability may be measured by the frequency, duration, and magnitude of adverse effects on the electric supply.” The council also suggests that reliability can be addressed by considering the two basic and functional aspects of the bulk electric system—adequacy and security. In this work, the discussion is focused on the adequacy aspect of power system reliability, which is defined as the static evaluation of the system’s ability to satisfy the system load requirements. In the context of the new business environment, market demand is interpreted as the system load. However, a secure implementation of electric power transactions concerns power system operation and stability issues:

1. Stability issue: The electric power system is a nonlinear dynamic system comprised of numerous machines synchronized with each other. Stable operation of these machines following disturbances or major changes in the network often requires limitations on various operating conditions, such as generation levels, load levels, and power transmission changes. Due to various inertial forces, these machines, together with other system components, require extra energy (reserve margins and load following capability) to safely and continuously actuate electric power transfer. 2. Thermal overload issue: Electrical network capacity and losses limit electric power transmission. Capacity may include real-time weather conditions as well as congestion management. The impact of transmission losses on market power is yet to be understood. 3. Operating voltage issues: Enough reactive power support must accompany the real power transfer to maintain the transfer capacity at the specified levels of open access.

In the new organizational structure, the services used for supporting a reliable delivery of electric energy (e.g., various reserve margins, load following capability, congestion management, transmission losses, reactive power support, etc.) are termed supportive services. These have been called “ancillary services” in the past. In this context, the term “ancillary services” is misleading since the services in question are not ancillary but closely bundled with the electric power transfer as described earlier. The open market system should consider all of these supportive services as an integral part of power transaction.

Planning Environments

15-5

This work proposes that supportive services become a competitive component in the energy market. It is embedded so that no matter what reasonable conditions occur, the (operationally) centralized service will have the obligation and the authority to deliver and keep the system responding according to adopted operating constraints. As such, although competitive, it is burdened by additional goals of ensuring reliability rather than open access only. The proposed pricing framework attempts to become economically efficient by moving from cost-based to price-based operation and introduces a mathematical framework to enable all players to be sufficiently informed in decision-making when serving other competitive energy market players, including customers.

15.2.3 Economic Evolution Some economists speculate that regional commodity exchanges within the United States would be oligopolistic in nature (having a limited numbers of sellers) due to the configuration of the transmission system. Some postulate that the number of sellers will be sufficient to achieve near-perfect competition. Other countries have established exchanges with as few as three players. However, such experiments have reinforced the notion that collusion is all too tempting, and that market power is the key to price determination, as it is in any other market. Regardless of the actual level of competition, companies that wish to survive in the deregulated marketplace must change the way they do business. They will need to develop bidding strategies for trading electricity via an exchange. Economists have developed theoretical results of how variably competitive markets are supposed to behave under varying numbers of sellers or buyers. The economic results are often valid only when aggregated across an entire industry and frequently require unrealistic assumptions. While considered sound in a macroscopic sense, these results may be less than helpful to a particular company (not fitting the industry profile) that is trying to develop a strategy that will allow it to remain competitive. GENCOs, energy service suppliers (ESSs), and DISTCOs that participate in an energy commodity exchange must learn to place effective bids in order to win energy contracts. Microeconomic theory states that in the long term, a hypothetical firm selling in a competitive market should price its product at its marginal cost of production. The theory is based on several assumptions (e.g., all market players will behave rationally, all market players have perfect information) that may tend to be true industrywide, but might not be true for a particular region or a particular firm. As shown in this work, the normal price offerings are based on average prices. Markets are very seldom perfect or in equilibrium. There is no doubt that deregulation in the power industry will have many far-reaching effects on the strategic planning of firms within the industry. One of the most interesting effects will be the optimal pricing and output strategies generator companies (GENCOs) will employ in order to be competitive while maximizing profits. This case study presents two very basic, yet effective means for a single generator company (GENCO) to determine the optimal output and price of their electrical power output for maximum profits. The first assumption made is that switching from a government regulated, monopolistic industry to a deregulated competitive industry will result in numerous geographic regions of oligopolies. The market will behave more like an oligopoly than a purely competitive market due to the increasing physical restrictions of transferring power over distances. This makes it practical for only a small number of GENCOs to service a given geographic region. The strongest force in the economic evolution is the demand for electricity. Customer\demand was assumed to be inelastic under many of the early market analysis studies. This ignored the success of demand side management in the regulated environment (Le et al., 1983; Cohen et al., 1987; Mortensen and Haggerty, 1990; Lee and Chen, 1992; Chu et al., 1993; Rupanagunta et al., 1995; Wei and Chen, 1995; Kurucz et al., 1996) and the cost reduction noticed by industrial and commercial customers (Bentley and Evelyn, 1987; Chen and Sheen 1993). Such asset flexibility is a large leverage in the management of the risk with portfolio management tools used in the competitive market. The drive to install smart grid components provide the infrastructure for demand side management as well as for distributed generation.

15-6

Power Systems Level 1 2 3 4

FERC

GENCO

SPUC

ESS/ESCO/EMCO

5

SPUC

..............

SPUC

NERC ICA/ISO/RTO TRANSCO

DISTCO

EMA

MARKCO

BROCO

FIGURE 15.2 Business environmental model.

15.2.4 Market Structure Although nobody knows the exact structure of the emerging deregulated industry, this research predicts that regional exchanges (i.e., electricity mercantile associations [EMAs]) will eventually play an important role. Electricity trading of the future will be accomplished through bilateral contracts and EMAs where traders bid for contracts via a double auction. The electric marketplace used in this section has been refined and described by various authors. Fahd and Sheblé (1992) demonstrated an auction mechanism. Sheblé (1994b) described the different types of commodity markets and their operation, outlining how each could be applied in the evolved electric energy marketplace. Sheblé and McCalley (1994) outlined how spot, forward, future, planning, and swap markets can handle real-time control of the system (e.g., automatic generation control) and risk management. Work by Kumar and Sheblé (1996b) brought the above ideas together and demonstrated a power system auction game designed to be a training tool. That game used the double auction mechanism in combination with classical optimization techniques. In several references (Kumar, 1996a, 1996b; Sheblé 1996; Richter 1997a), a framework is described in which electric energy is only sold to DISTCOs, and electricity is generated by GENCOs (see Figure 15.2). The NERC sets the reliability standards. Along with DISTCOs and GENCOs, ESCOs, ancillary services companies (ANCILCOs), and TRANSCOs interact via contracts. The contract prices are determined through a double auction. Buyers and sellers of electricity make bids and offers that are matched subject to approval of the independent contract administrator (ICA), who ensures that the contracts will result in a system operating safely within limits. The ICA submits information to an ISO for implementation. The ISO is responsible for physically controlling the system to maintain its security and reliability. As of this writing, the ISO has integrated the functions of EMA, MARKCO, and BROCO within itself. The ESS provide basic or extended services to the customers. ESCOs have provided minimal contracts simply to provide energy. EMCOs provide contracts and services to reduce the costs of energy. The advent of the smart grid enables customers to participate on the hourly transactions and pay by time of use instead of average price. Market companies (MARKCO) enable anyone to invest in the energy markets, especially speculators. Speculators already exist at most trading companies but the general public is not able to participate as in the other commodity markets. Until the general public is enabled to enter the markets, brokerage companies (BROCOs) are not needed.

15.2.5 Fully Evolved Marketplace The following sections outline the role of a horizontally integrated industry. Many curious acronyms have described generation companies (IPP, QF, Cogen, etc.), transmission companies (IOUTS, NUTS, etc.), and distribution companies (IOUDC, COOPS, MUNIES, etc.). The acronyms used in this work are described in the following sections. 15.2.5.1 Horizontally Integrated The restructuring of the electric power industry is most easily visualized as a horizontally integrated marketplace. This implies that interrelationships exist between generation (GENCO), transmission (TRANSCO), and distribution (DISTCO) companies as separate entities. Note that independent power producers (IPP),

Planning Environments

15-7

qualifying facilities (QF), etc. may be considered as equivalent generation companies. Nonutility transmission systems (NUTS) may be considered as equivalent transmission companies. Cooperatives and municipal utilities may be considered as equivalent distribution companies. All companies are assumed to be coordinated through a regional Transmission Corporation (or regional transmission group). 15.2.5.2 Federal Energy Regulatory Commission (FERC) FERC is concerned with the overall operation and planning of the national grid, consistent with the various energy acts and public utility laws passed by Congress. Similar federal commissions exist in other government structures. The goal is to provide a workable business environment while protecting the economy, the customers, and the companies from unfair business practices and from criminal behavior. GENCOs, ESCOs, and TRANSCOs would be under the jurisdiction of FERC for all contracts impacting interstate trade. 15.2.5.3 State Public Utility Commission (SPUC) SPUCs protect the individual state economies and customers from unfair business practices and from criminal behavior. It is assumed that most DISTCOs would still be regulated by SPUCs under performancebased regulation and not by FERC. GENCOs, ESCOs, and TRANSCOs would be under the jurisdiction of SPUCs for all contracts impacting intrastate trade. 15.2.5.4 Generation Company (GENCO) The goal for a generation company, which has to fill contracts for the cash and futures markets, is to package production at an attractive price and time schedule. One proposed method is similar to the classic decentralization techniques used by a vertically integrated company. The traditional power system approach is to use Dantzig–Wolfe decomposition. Such a proposed method may be compared with traditional operational research methods used by commercial market companies for a “make or buy” decision. 15.2.5.5 Transmission Company (TRANSCO) The goal for transmission companies, which have to provide services by contracts, is to package the availability and the cost of the integrated transportation network to facilitate transportation from suppliers GENCOs to buyer ESCOs. One proposed method is similar to oil pipeline networks and energy modeling. Such a proposed method can be compared to traditional network approaches using optimal power flow programs. 15.2.5.6 Distribution Company (DISTCO) The goal for distribution companies, which have to provide services by contracts, is to package the availability and the cost of the radial transportation network to facilitate transportation from suppliers GENCOs to buyers ESCOs. One proposed method is similar to distribution outlets. Such proposed methods can be compared to traditional network approaches using optimal power flow programs. The disaggregation of the transmission and the distribution system may not be necessary, as both are expected to be regulated as monopolies at the present time. 15.2.5.7 Energy Service Supplier (ESS), Energy Service Company (ESCO), Energy Management Company (EMCO) A primary tool of the ESS company is to implement demand side management (DSM) to control demand not only to reduce the peak demand but also to provide flexibility for risk management, for alternative ancillary service sources, and for other more complex risk and service enterprises (Daryanian et al., 1989; Ng and Sheblé, 1998; Lee and Chen, 1992, 1993, 1994). A secondary tool is the availability of distributed generation with natural gas micro turbines, wind generators, and solar cells. The advent of more biofuels will provide more flexibility of hybrid diesel automobiles as distributed generation. The eventual implementation of the hydrogen economy as methane gas (hydrogen combined with CO2) will

15-8

Power Systems

extend the use of distributed generation with cost-effective storage of CH4 for micro turbines and hybrid automobiles (Sheblé, 1999a,b). The goal for energy service companies, which may be large industrial customers or customer pools, is to purchase power at the least cost when needed by consumers. One proposed method is similar to the decision of a retailer to select the brand names for products being offered to the public. Such a proposed method may be compared to other retail outlet shops. 15.2.5.8 Independent System Operator (ISO) The primary concern is the management of operations. Real-time control (or nearly real-time) must be completely secure if any amount of scheduling is to be implemented by markets. The present business environment uses a fixed combination of units for a given load level, and then performs extensive analysis of the operation of the system. If markets determine schedules, then the unit schedules may not be fixed sufficiently ahead of realtime for all of the proper analysis to be completed by the ISO. 15.2.5.9 Regional Transmission Organization (RTO) The goal for a regional transmission group, which must coordinate all contracts and bids among the three major types of players, is to facilitate transactions while maintaining system planning. One proposed method is based on discrete analysis of a Dutch auction. Other auction mechanisms may be suggested. Such proposed methods are similar to a warehousing decision on how much to inventory for a future period. As shown later in this work, the functions of the RTG and the ISO could be merged. Indeed, this should be the case based on organizational behavior. 15.2.5.10 Independent Contract Administrator (ICA) The goal for an ICA is a combination of the goals for an ISO and an RTG. Northern States Power Company originally proposed this term. This term will be used in place of ISO and RTG in the following to differentiate the combined responsibility from the existing ISO companies. 15.2.5.11 Energy Mercantile Association (EMA), Market Company (MARKCO), Brokerage Company (BROCO) Competition may be enhanced through the various markets: cash, futures, planning, and swap. The cash market facilitates trading in spot and forward contracts. This work assumes that such trading would be on an hourly basis. Functionally, this is equivalent to the interchange brokerage systems implemented in several states. The distinction is that future time period interchange (forward contracts) are also traded. The futures market facilitates trading of futures and options. These are financially derived contracts used to spread risk. The planning market facilitates trading of contracts for system expansion. Such a market has been proposed by a west coast electric utility. The swap market facilitates trading between all markets when conversion from one type of contract to another is desired. It should be noted that multiple markets are required to enable competition between markets. The structure of any spot market auction must include the ability to schedule as far into the future as the industrial practice did before deregulation. This would require extending the spot into the future for at least 6 months, as proposed by this author (Sheblé, 1994). Future month production should be traded for actual delivery in forward markets. Future contracts should be implemented at least 18 months into the future if not 3 years. Planning contracts must be implemented for at least 20 years into the future, as recently offered by TVA, to provide an orderly, predictable expansion of the generation and transmission systems. Only then can timely addition of generation and transmission be assured. Finally, a swap market must be established to enable the transfer of contracts from one period (market) to another. To minimize risk, the use of option contracts for each market should be implemented. Essentially, all of the players share the risk. This is why all markets should be open to the public for general trading and subject to all rules and regulations of a commodity exchange. Private exchanges, not subject to such regulations, do not encourage competition and open price discovery.

15-9

Planning Environments Time horizon (Months) 0 ....................1 Spot market

2 ............. 12 Forward market

.................

18

.................

360

Swap market (Market to market contracts) Futures market

Planning market

FIGURE 15.3 Interconnection between markets.

The described framework (Sheblé, 1996) allows for cash (spot and forward), futures, and planning markets as shown in Figure 15.3. The spot market is most familiar within the electric industry (Schweppe et al. 1988). A seller and a buyer agree (either bilaterally or through an exchange) upon a price for a certain amount of power (MW) to be delivered sometime in the near future (e.g., 10 MW from 1:00 p.m. to 4:00 p.m. tomorrow). The buyer needs the electricity, and the seller wants to sell. They arrange for the electrons to flow through the electrical transmission system. A forward contract is a binding agreement in which the seller agrees to deliver an amount of a particular product in a specified quality at a specified time to the buyer. A futures contract is further into the future than is the spot market. In both the forward and spot contracts, the buyer and seller want physical goods (e.g., the electrons). A futures contract is primarily a financial instrument that allows traders to lock in a price for a commodity in some future month. This helps traders manage their risk by limiting potential losses or gains. Futures contracts exist for commodities in which there is sufficient interest and in which the goods are generic enough that it is not possible to tell one unit of the good from another (e.g., 1 MW of electricity of a certain quality, voltage level, etc.). A futures option contract is a form of insurance that gives the option purchaser the right, but not the obligation, to buy (sell) a futures contract at a given price. For each options contract, there is someone “writing” the contract who, in return for a premium, is obligated to sell (buy) at the strike price (see Figure 15.3). Both the options and the futures contracts are financial instruments designed to minimize risk. Although provisions for delivery exist, they are not convenient (i.e., the delivery point is not located where you want it to be located). The trader ultimately cancels his position in the futures market, either with a gain or loss. The physicals are then purchased on the spot market to meet demand with the profit or loss having been locked in via the futures contract. A swap is a customized agreement in which one firm agrees to trade its coupon payment for one held by another firm involved in the swap. Finally, a planning market is needed to establish a basis for financing long term projects like transmission lines and power plants (Sheblé, 1993).

15.2.6 Computerized Auction Market Structure Auction market structure is a computerized market, as shown in Figure 15.4. Each of the agents has a terminal (PC, workstation, etc.) connected to an auctioneer (auction mechanism) and a contract evaluator. Players generate bids (buy and sell) and submit the quotation to the auctioneer. A bid is a specified amount of electricity at a given price. The auctioneer binds bids (matching buyers and sellers) subject to approval of the contract evaluation. This is equivalent to the pool operating convention used in the vertically integrated business environment. ICA

Contract evaluations

Auction mechanism Communication player to auction Player

Player

FIGURE 15.4 Computerized markets.

Player

Player

15-10

Power Systems Time line into future (hours) Real time.........................................................................................Planning horizon MARKCO GENCO ESCO BROCO Bids Bids ...................... Bids Futures Forward Planning Spot Swap GENCO

ESCO

...

...

ICA/ISO/RTO MARKCO ...

Information Level Player Communication Markets Market

BROCO

Coordination Players

...

...

FIGURE 15.5 Electric market.

The contract evaluator verifies that the network can remain in operation with the new bid in place. If the network cannot operate, then the match is denied. The auctioneer processes all bids to determine which matches can be made. However, the primary problem is the complete specification of how the network can operate and how the agents are treated comparably as the network is operated closer to limits. The network model must include all constraints for adequacy and security. The major trading objectives are hedging, speculation, and arbitrage. Hedging is a defense mechanism against loss and/or supply shortages. Speculation is assuming an investment risk with a chance for profit. Arbitrage is crossing sales (purchases) between markets for riskless profit. This work assumes that there are four markets commonly operated: forward, futures, planning, and swaps (Figure 15.5). Forward Market: The forward contracts reflect short term future system conditions. In the forward market, prices are determined at the time of the contract but the transactions occur at some future time. Optimization tools for short term scheduling problems can be enhanced to evaluate trading opportunities in the forward market. For example, short term dispatching algorithms, such as economic unit commitment dispatch, can be used to estimate and earn profit in the forward market. Futures Market: A futures market creates competition because it unifies diverse and scattered local markets and stabilizes prices. The contracts in the futures market are risky because price movements over time can result in large gains or losses. There is a link between forward markets and futures markets that restricts price volatility. Options (options contracts) allow the agent to exercise the right to activate a contract or cancel it. Claims to buy are called “call” options. Claims to sell are called “put” options. A more detailed discussion of an electric futures contract is discussed in Sheblé (1994b). The components include trading unit, trading hours, trading months, price quotation, minimum price fluctuation, maximum daily price fluctuation, last trading day, exercise of options, option strike prices, delivery, delivery period, alternate delivery procedure, exchange of futures for, or in connection with, physicals, quality specifications, and customer margin requirements. The implementation proposals include optimization techniques at the hourly level (or faster) with the unit commitment selection under the GENCO decision domain (Fahd and Sheblé, 1992; Fahd et al., 1992; Smith, 1993; Post et al., 1995) and integrated unit commitment models with optimal power flow dispatches for each hour (Kumar and Sheblé, 1994, 1996a,b; Schweppe et al., 1988; Dekrajangpetch and Sheblé, 1998). Swap Market: In the swap market, contract position can be closed with an exchange of physical or financial substitutions. The trader can find another trader who will accept (make) delivery and end the trader’s delivery obligation. The acceptor of the obligation is compensated through a price discount or a premium relative to the market rate. The financial drain inflicted on traders when hedging their operations in the futures market is slightly higher than the one inflicted through direct placement in the forward market. An optimal mix of options, forward commitments, futures contracts, and physical inventories is difficult to assess and

Planning Environments

15-11

depends on hedging, constraints imposed by different contracts, and the cost of different contracts. A clearinghouse such as a swap market handles the exchange of various energy instruments. Planning Market: The growth of transmission grid requires transmission companies to make contracts based on the expected usage to finance projects. The planning market would underwrite equipment usage subject to the long term commitments to which all companies are bound by the rules of network expansion to maintain a fair marketplace. The network expansion would have to be done to maximize the use of transmission grid for all agents. Collaboration would have to be overseen and prohibited with a sufficiently high financial penalty. The growth of the generation supply similarly requires such markets. However, such a market has been started with the use of franchise rights (options) as established in recent Tennessee Valley Authority connection contracts. This author has published several papers outlining the need for such a market. Such efforts are not documented in this work.

15.2.7 Capacity Expansion Problem Definition The capacity expansion problem is different for an ESCO, GENCO, TRANSCO, DISTCO, and ANSILCO. This section assumes that the ICA will not own equipment but will only administer the contracts between players. The capacity expansion problem is divided into the following areas: generation expansion, transmission expansion, distribution expansion, and market expansion. ESCOs are concerned with market expansion. GENCOs are concerned with generation expansion. TRANSCOs are concerned with transmission expansion. DISTCOs are concerned with distribution expansion. ANSILCOs are concerned with supportive devices expansion. This author views ancillary services as a misnomer. Such services are necessary supportive services. Thus, the term “supportive” will be used instead of ancillary. Also, since supportive devices are inherently part and parcel of the transmission or distribution system, these devices will be assumed into the TRANSCO and DISTCO functions without loss of generality. Thus, ANSILCOs are not treated separately. Based on the above idealized view of the marketplace, the following generalizations are made. GENCOs are concerned with the addition of capacity to meet market demands while maximizing profit. Market demands include bilateral contracts with the EMA as well as bilateral contracts with ESCOs or with the ICA. ESCOs are concerned with the addition of capacity of supplying customers with the service desired to maintain market share. ESCOs are thus primarily concerned with the processing of information from marketplace to customer. However, ESCOs are also concerned with additional equipment supplied by DISTCOs or TRANSCOs to provide the level of service required by some customers. ESCOs are thus concerned with all aspects of customer contracts and not just the supply of “electrons.” The ICA is concerned with the operation of the overall system subject to the contracts between the buyers and the sellers and between all players with ICA. The overall goal of the ICA is to enable any customer to trade with any other customer with the quick resolution of contract enforcement available through mercantile associations. The ICA maintains the reliability of the network by resolving the unexpected differences between the contracts, real operation, and unplanned events. The ICA has the authority, through contracts, to buy generation services, supportive services, and/or transmission services, or to curtail contracts if the problems cannot be resolved with such purchases as defined in these contracts. Thus, the ICA has the authority to connect or disconnect generation and demand to protect the integrity of the system. The ICA has the authority to order new transmission or distribution expansion to maintain the system reliability and economic efficiency of the overall system. The economic efficiency is determined by the price of electricity in the cash markets on a periodic basis. If the prices are approximately the same at all points in the network, then the network is not preventing customers from getting to the suppliers. Similarly, the suppliers can get to the buyers. Since all buyers and suppliers are protected from each other through the default clauses of the mercantile agreement, it does not matter which company deals with other companies as the quick resolution of disputes is guaranteed.

15-12

Power Systems

This strictness of guarantee is the cornerstone of removing the financial uncertainty at the price of a transaction fee to cover the costs of enforcement. The goal of each company is different but the tools are the same for each. First, the demand must be predicted for future time periods sufficiently into the future to maintain operation financially and physically. Second, the present worth of the expansion projects has to be estimated. Third, the risks associated with each project and the demand-forecast uncertainty must be estimated. Fourth, the acceptable value at risk acceptable for the company has to be defined. Fifth, the value at risk has to be calculated. Sixth, methods of reducing the value at risk have to be identified and evaluated for benefits. Seventh, the overall portfolio of projects, contracts, strategies, and risk has to be assessed. Only then can management decide to select a project for implementation. The characteristics of expansion problems include

1. The cost of equipment or facilities should exhibit economies of scale for the same risk level baring technology changes. 2. Time is a primary factor since equipment has to be in place and ready to serve the needs as they arise. Premature installation results in idle equipment. Delayed installation results in lost market share. 3. The risk associated with the portfolio of projects should decrease as time advances. 4. The portfolio has to be revalued at each point when new information is available that may change the project selection, change the strategy, or change the mix of contracts.

The capital expansion problem is often referred to as the “capital budgeting under uncertainty” problem (Aggarwal, 1993). Thus, capital expansion is an exercise in estimating the present net value of future cash flows and other benefits as compared to the initial investment required for the project given the risk associated with the project(s). The key concept is the uncertainty and thus the risk of all business ventures. Uncertainties may be due to estimation (forecasting) and measurement errors. Such uncertainties can be reduced by the proper application of better tools. Another approach is to investment in information technology to coordinate the dissemination of information. Indeed, information technology is one key to the appropriate application of capital expansion. Another uncertainty factor is that the net present value depends on market imperfections. Market imperfections are due to competitor reactions to each other’s strategies, technology changes, and market rule changes (regulatory changes). The options offered by new investment are very hard to forecast. Also the variances of the options to reduce the risk of projects are critical to proper selection of the right project. Management has to constantly revalue the project, change the project (including termination), integrate new information, or modify the project to include technology changes. Estimates have often been biased by management pressure to move ahead, to not investigate all risks, or to maintain strategies that are not working as planned. Uncertainties in regulations and taxes are often critical for the decision to continue. There are three steps to any investment plan: investment alternative identification, assessment, selection and management of the investment as events warrant. Capacity expansion is one aspect of capital budgeting. Marketing and financial investments are also capital budgeting problems. Often, the capacity expansion has to be evaluated not only on the projects merits, but also the merits of the financing bundled with the project.

15.3 Regulated Environment The regulated industry is structured as a vertically integrated business model. This organization is typified by the structure shown in Figure 15.6. This structure should be compared with the previously defined competitive business environmental model shown in Figure 15.2. The entities previously described function in the same manner in this business model. The difference is that the utility company is an umbrella company or an integrated company with divisions segmented by type of expertise

15-13

Planning Environments

Level 1 2 3 4 5

FERC

SPUC

NERC UTILITY GENCO/ESS/TRANSCO/DISTCO NETRELCO

FIGURE 15.6 Regulated business environmental model.

and equipment. All divisions report to the utility and take direction from the utility in all business operations. The utility seeks tariff approval from the state public utility commission and the federal energy regulatory commission. The interchange between utility companies is approved by both FERC and SPUCs. The interchange capabilities and contract approvals are handled by the network reliability management company (NETRELCO), which coordinates network activities between utilities. Alternative umbrella company structures are in use within the United States. American Electric Power and Southern Company are umbrella companies that own several utilities in several different states within the United States. Due to SPUC regulations and tariff approvals, utilities are located within a state boundary within the United States for ease of accounting and of regulation. The Mid-America Interpool Network (MAIN) Company was one such network reliability management company that coordinated maintenance and interchange schedules between utilities in the Midwestern part of the United States. MAIN was announced on November 24, 1964, by American Electric Power, Commonwealth Edison, the Illinois-Missouri Pool, the Indiana Power Pool, and the Wisconsin Planning Group. It later became Mid-America Interconnected Network, Inc. and was the first regional electric reliability council. When the North American Electric Reliability Council (NERC) was later formed, MAIN became 1 of the 10 electric reliability councils that comprised NERC. MAIN served the electric utilities in the Midwest for over 41 years. The need for a NETRELCO was due to the large amount of interchange between utilities. Interchange is beneficial to lower costs each hour (economy a), to lower costs over a number of sequential hours (economy b), and to increase reliability with more inertia and capacity as a response to disturbances (Wood and Wollenberg, 1996). The types of interchange grew quickly justifying more transmission between companies for short- and long-term economic reasons as well as for reliability reasons (Kelley et al., 1987; Winston and Gibson, 1988; Parker et al., 1989; Rau, 1989; Shirmohammadi et al., 1989; Clayton et al., 1990; Happ, 1990; Svoboda and Oren, 1994; Tabors, 1994; Post et al., 1995; Vojdani et al., 1995). Interchange on the east cost of the United States led to pool operation as demonstrated by the formation of the PJM power pool. An alternative form was the pool company (POOLCO) such as the Pennsylvania-New JerseyMaryland (PJM) power pool. PJM was formed in 1927 when three utilities, realizing the benefits and efficiencies possible by interconnecting to share their generating resources as if one utility, formed the world’s first continuing power pool. The power pool operated a joint dispatch amongst the utilities for continuous economy, a type of interchange that was operated as if by one company [PJM]. Additional utilities joined in 1956, 1965, and 1981. Throughout this time, PJM was operated by a department of one member utility with contractual interchange agreements to dynamically alter the amounts of interchange and the costs of interchange amongst the participants. In 1962, PJM installed its first online computer to control generation and dispatch generation with dynamic interchange between the member utility companies. PJM broadcasted a system lambda that each unit used as the economic solution to the overall dispatch of all companies. PJM completed its first energy management system (EMS) in 1968 and included transfer limitations between the member companies using a linear network solution technique. The EMS is the information technology system that makes it possible to monitor transmission grid operations in real time as a NETRELCO. The primary task of a NETRELCO was to study the reliability of the electric power system (Breiphol, 1990;

15-14

Power Systems

Billinton and Lian, 1991, Billinton and Wenyuan, 1991, Billinton and Gan, 1993, Billinton and Li, 1994). In 1996, PJM launched its first website to provide its members with current system information. PJM transitioned to an ISO in 1997. PJM includes the functions of an ICA, EMA, NETRELCO, and RTO in the competitive environment. The transition of PJM from regulated to competitive market environment is a study of the subtle changes between the two environments. Dynamic interchange was extended for the joint operation of jointly owned units. Planning solutions in the Midwest section of the United States found that sharing the cost of unit construction and operation was economic (Podmore et al., 1979; Lee, 1988). Such units were called jointly owned units (JOUs) and the interchange between member owners became dynamically dispatched so each member could use the unit for economic operation continuously. The resulting dynamic alteration of a unit’s output led to dynamic interchange between the members, mimicking a power pool with the JOUs. The ultimate JOU in the United States is the Intermountain Power Project (IPP). This JOU installation is owned by over 50 companies. Some of the companies receive the output through dynamic interchange agreements on the Utah-based alternating current (AC) transmission system, others receive the output across the high-voltage direct current (HVDC) link to California. The Florida Energy Broker was another implementation of a pool type operation for continuous (hourly) economic dispatch of generation within the state of Florida. This organization collected unit bids and asks for auction matching in the 1980s. The resulting matches were implemented by each utility each hour resulting in significant operational savings. As such, this was the first implementation of a competitive environment using an auction mechanism (Cohen, 1982). That implementation was the basis for auction market simulation (Sheblé, 1994a,b). The planning of utilities under such a business environment has been well documented by general textbooks and papers (Adams et al., 1972; Booth, 1972; Chao, 1983; Merril, 1991; Merril and Wood, 1990; Sullivan, 1977; Wang and McDonald, 1994; Willis, 1996; Seifi and Sepasian, 2001; Li, 2011). The production costing of generation was one of the first computerized planning tools used in both business environments (Baleriaux, 1967; Day, 1971). The application of probabilistic production costing has been one of the most thoroughly researched areas (Lee, F, 1988; Lin et al., 1989; Wang, 1989; Billinton and Lian, 1991; Billinton and Wenyuan, 1991; Billington and Li, 1992; Billington and Gan, 1993; Delson et al., 1991; Huang and Chen, 1993; Pereiran et al., 1992; Lee and Chen, 1992, 1993, 1994; Miranda, 1994; Parker and Stremel, 1996). The other area of intense research is in the area of transaction selection between utilities (Kelley et al., 1987; Winston and Gibson, 1988; Parker et al., 1989; Rau, 1989; Clayton et al., 1990; Happ, 1990; Shirmohammadi et al., 1989; Roy, 1993). The mixed business models used within the United States has spawned a number of mixed tools to provide the benefits of both environments (David and Li, 1993; Krause and McCalley, 1994; Rakic and Markovic, 1994; Post et al., 1995; Vojdani et al., 1995; Wu and Varaiya, 1995; Clayton and Mukerji, 1996; Richter and Sheblé, 1997a, 1997b, 1998). The application of competitive markets compared to traditional regulation has been analyzed but more work is required (Hobbs and Schuler, 1985; Hogan, 1977, Lerner, 1994; O’Neill and Whitmore, 1994; Oren et al., 1994; Oren, 1997; Smith, 1988; Bhattacharya et al., 2001; Loi, 2001).

15.4 Other Sections on Planning The following sections on planning deal with the overall approach as described by Dr. H. Merrill and include sections on forecasting, power system planning, transmission planning, and system reliability. Forecasting demand is a key issue for any business entity. Forecasting for a competitive industry is more critical than for a regulated industry. Transmission planning is discussed based on probabilistic techniques to evaluate the expected advantages and costs of present and future expansion plans. Reliability of the supply is covered, including transmission reliability. The most interesting aspect of the electric power industry is the massive changes presently occurring. It will be interesting to watch as the industry adapts to regulatory changes and as the various market players find their corporate niche in this new framework.

Planning Environments

15-15

References Adams, N., F. Belgari, M.A. Laughton, and G. Mitra, Mathematical programming systems in generation, transmission and distribution planning, PSCC Proceedings Paper Number 1, 1/13, Grenoble, France, 1972. Aggarwal, R., Capital Budgeting Under Uncertainty, Prentice-Hall, Englewood Cliffs, NJ, 1993. Baleriaux, H., Simulation de l’exploitation d’un parc de machines thermiques de production d’electricite couple a des stations de pompage, Revue E, 5(7), 1–24, 1967. Baughman, M.L., J.W. Jones, and A. Jacob, Model for evaluating the economics of cool storage systems, IEEE Trans. Power Syst., 8(2), 716–722, May 1993. Bently, W.G. and J.C. Evelyn, Customer thermal energy storage: A marketing opportunity for cooling off electric peak demand, IEEE Trans. Power Syst., 1(4), 973–979, 1987. Bhattacharya, K., M.H.J. Bollen, and J.E. Daalder, Operation of Restructured Power Systems, Springer, New York, 2001. Billinton, R. and L. Gan, Monte Carlo simulation model for multiarea generation system reliability studies, IEE Proc., 140(6), 532–538, 1993. Billinton, R. and W. Li, A Monte Carlo method for multi-area generation system reliability assessment, IEEE Trans. Power Syst., 7(4), 1487–1492, 1992. Billinton, R. and W. Li, Reliability Assessment of Electrical Power Systems Using Monte Carlo Methods, Springer, New York, 1994. Billinton, R. and G. Lian, Monte Carlo approach to substation reliability evaluation, IEE Proc., 140(2), 147–152, 1991. Billinton, R. and L. Wenyuan, Hybrid approach for reliability evaluation of composite generation and transmission systems using Monte-Carlo simulation and enumeration technique, IEE Proc., 138(3), 233–241, 1991. Binger, B.R. and E. Hoffman, Microeconomics with Calculus, Scott, Foresman and Company, Glenview, IL, 1988. Booth, R.R., Optimal generation planning considering uncertainty, IEEE Trans. PAS-91, 70–77, Jan./Feb. 1972. Breipohl, A.M., Probabilistic Systems Analysis: An Introduction to Probabilistic Models, Decisions, and Applications of Random Processes, Wiley, New York, 1990. Bussey, L. E., The Economic Analysis of Industrial Projects, Prentice-Hall, Englewood Cliffs, NJ, 1981. Chao, H.P., Peak load pricing and capacity planning with demand and supply uncertainty, Bell J. Econ., 14(1), 179–190, 1983. Chen, C.S. and J.N. Sheen, Cost benefit analysis of a cooling energy storage system, IEEE Trans. Power Syst., 8(4), 1504–1510, 1993. Chu, W., B. Chen, and C. Fu, Scheduling of direct load control to minimize load reduction for a utility suffering from generation shortage, IEEE/PES Winter Meeting, Columbus, OH, 1993. Clayton, J.S., S.R. Erwin, and C.A. Gibson, Interchange costing and wheeling loss evaluation by means of incrementals, IEEE Trans. Power Syst., 5(3), 759–765, 1990. Clayton, R.E. and R. Mukerji, System planning tools for the competitive market, IEEE Comput. Appl. Power, 9(3), 50–55, 1996. Cohen, L., A Spot Market for Electricity: Preliminary Analysis of the Florida Energy Broker, RAND Note, N-1817-DOE, February 1982. Cohen, A.I., J.W. Patmore, D.H. Oglevee, R.W. Berman, L.H. Ayers, and J.F. Howard, An integrated system for load control, IEEE Trans. Power Syst., PWRS-2(3), 645–651, 1987. Daellenbach, H.G., Systems and Decision Making, John Wiley & Sons, New York, 1994. Daryanian, B., R.E. Bohn, and R.D. Tabors, Optimal demand-side response to electricity spot prices for storage-type customers, IEEE Trans. Power Syst., 4(3), 897–903, 1989.

15-16

Power Systems

David, A.K. and Y.Z. Li, Effect of inter-temporal factors on the real-time pricing of elasticity, IEEE Trans. Power Syst., 8(1), 44–52, 1993. Day, J.T., Forecasting minimum production costs with linear programming, IEEE Trans. Power Appar. Syst., PAS-90(2), 814–823, 1971. Dekrajangpetch, S. and G.B. Sheblé, Alternative implementations of electric power auctions, in Proceedings of the 60th American Power Conference, Vol. 60-1, pp. 394–398, 1998. Delson, J.K., X. Feng, and W.C. Smith, A validation process for probabilistic production costing programs, IEEE Trans. Power Syst., 6(3), 1326–1336, 1991. Fahd, G., Optimal power flow emulation of interchange brokerage systems using linear programming, PhD dissertation, Auburn University, Auburn, AL, 1992. Fahd, G., D.A. Richards, and G.B. Sheblé, The implementation of an energy brokerage system using linear programming, IEEE Trans. Power Syst., T-PWRS, 7(1), 90–96, 1992. Fahd, G. and G. Sheblé, Optimal power flow of interchange brokerage system using linear programming, IEEE Trans. Power Syst., T-PWRS, 7(2), 497–504, 1992. Happ, H.H., Report on Wheeling Costs, Case 88-E-238, The New York Public Service Commission, February 1990. Hobbs, B.F. and R.E. Schuler, An assessment of the deregulation of electric power generation using network models of imperfect spatial markets, Papers of the Regional Science Association, 57(1), 75–89, 1985. Hogan, W.W., A market power model with strategic interaction in electricity networks, Energy J., 18(4), 107–141, 1997. Huang, S.R. and S.L. Chen, Evaluation and improvement of variance reduction in Monte-Carlo production simulation, IEEE Trans. Energy Convers., 8(4), 610–619, 1993. Ilic, M., F. Galiana, and L. Fink, Power Systems Restructuring: Engineering and Economics, Kluwer Academic Publishers, Norwell, MA, 1998. Kelley, K., S. Henderson, P. Nagler, and M. Eifert, Some Economic Principles for Pricing Wheeled Power, National Regulatory Research Institute, Silver Spring, MD, August 1987. Krause, B.A. and J. McCalley, Bulk power transaction selection in a competitive electric energy system with provision of security incentives, Proceedings of the 26th Annual North American Power Symposium, Manhattan, KS, pp. 126–136, September 1994. Kumar, J., Electric power auction market implementation and simulation, PhD dissertation, Iowa State University, Ames, IA, 1996. Kumar, J. and G.B. Sheblé, A framework for transaction selection using decision analysis based upon risk and cost of insurance, Proceedings of the 29th North American Power Symposium, Kansas State University, Manhattan, KS, pp. 548–557, 1994. Kumar, J. and G.B. Sheblé, A decision analysis approach to transaction selection problem in a competitive electric market, Electric Power Syst. Res. J., 38(3), 209–216, 1996a. Kumar, J. and G.B. Sheblé, Transaction selection using decision analysis based upon risk and cost of insurance, IEEE Winter Power Meeting, Baltimore, MD, 1996b. Kurucz, C.N., D. Brandt, and S. Sim, A linear programming model for reducing system peak through customer load control programs, IEEE PES Winter Meeting, 96 WM 239–9 PWRS, Baltimore, MD, 1996. Lai, L.L., Power System Restructuring and Deregulation: Trading, Performance and Information Technology, John Wiley & Sons, Chichester, U.K., 2001. Le, K.D., R.F. Boyle, M.D. Hunter, and K.D. Jones, A procedure for coordinating direct-load-control strategies to minimize system production cost, IEEE Trans. Power Appar. Syst., PAS-102(6), 1983. Lee, F.N., Three-area joint dispatch production costing, IEEE Trans. Power Syst., 3(1), 294–300, 1988. Lee, T.Y. and N. Chen, The effect of pumped storage and battery energy storage systems on hydrothermal generation coordination, IEEE Trans. Energy Convers., 7(4), 631–637, 1992. Lee, T.Y. and N. Chen, Optimal capacity of the battery storage system in a power system, IEEE Trans. Energy Convers., 8(4), 667–673, 1993.

Planning Environments

15-17

Lee, T.Y. and N. Chen, Effect of battery energy storage system on the time-of-use rates industrial customers, IEE Proc. Gener. Transm. Distrib., 141(5), 521–528, 1994. Lee, S.H. and C.L. Wilkins, A practical approach to appliance load control analysis: A water heater case study, IEEE Trans. Power Appar. Syst., 7(4), 1992. Lerner, A.P., Monopoly and the measurement of monopoly power, Rev. Econ. Stud., 1, 157–175, 1934. Li, W., Probabilistic Transmission System Planning, IEEE Press, Hoboken, NJ, 2011. Lin, M., A. Breipohl, and F. Lee, Comparison of probabilistic production cost simulation methods, IEEE Trans. Power Syst., 4(4), 1326–1333, 1989. McCalley, J., A. Fouad, V. Vittal, A. Irizarry-Rivera, R. Farmer, and B. Agarwal, A probabilistic problem in electric power system operation: The economy-security tradeoff for stability-limited systems, Proceedings of the Third International Workshop on Rough Sets and Soft Computing, November 10–12, 1994, San Jose, CA. McCalley, J. and G.B. Sheblé, Competitive electric energy systems: Reliability of bulk transmission and supply, tutorial paper presented at the Fourth International Conference of Probabilistic Methods Applied to Power Systems, Brazil, 1994. Merrill, H.M., Have I Ever Got a Deal for You. Economic Principles in Pricing of Services, IEEE SP 91EH0345-9-PWR, pp. 1–8, 1991. Merril, H.M. and A.J. Wood, Risk and uncertainty in power system planning, 10th Power Systems Computation Conference, PSCC, Graz, Austria, August 1990. Miranda, V., Power system planning and fuzzy sets: Towards a comprehensive model including all types of uncertainties, Proceedings of the PMAPSí94, Rio de Janeiro, Brazil, September 1994. Miranda, V. and L.M. Proença, A general methodology for distribution planning under uncertainty, including genetic algorithms and fuzzy models in a multi-criteria environment, Proceedings of Stockholm Power Technology, SPT’95, Stockholm, Sweden, June 18–22, pp. 832–837, 1995. Mortensen, R.E. and K.P. Haggerty, Dynamics of heating and cooling loads: Models, simulation, and actual utility data, IEEE Trans. Power Syst., 5(1), 253–248, 1990. Ng, K.-H., Reformulating load management under deregulation, Master’s thesis, Iowa State University, Ames, IA, May 1997. Ng, K.-H. and G.B. Sheblé, Direct load control—A profit-based load management using linear programming, IEEE Trans. Power Syst., 13(2), 688–694, 1998. O’Neill, R.P. and C.S. Whitmore, Network oligopoly regulation: An approach to electric federalism, Electr. Federalism Symp., June 24, 1993 (Revised March 16, 1994). Oren, S.S., Economic inefficiency of passive transmission rights in congested electricity systems with competitive generation, Energy J., 18(1), 63–83, 1997. Oren, S.S., P. Spiller, P. Varaiya, and F. Wu, Nodal prices and transmission rights: A critical appraisal, University of California at Berkeley Research Report, December 1994. Outhred, H.R., Principles of a market-based electricity industry and possible steps toward implementation in Australia, International Conference on Advanced Power System Control, Operation and Management, Hong Kong, December 7–10, 1993. Parker, B.J., E. Denzinger, B. Porretta, G.J. Anders, and M.S. Mirsky, Optimal economic power transfers, IEEE Trans. Power Syst., 4(3), 1167–1175, 1989. Parker, C. and J. Stremel, A smart Monte Carlo procedure for production costing and uncertainty analysis, Proceedings of the American Power Conference, 58(II), 897–900, 1996. Pereira, V., B.G. Gorenstin, and M. Morozowski Fo, Chronological probabilistic production costing and wheeling calculations with transmission network modeling, IEEE Trans. Power Syst., 7(2), 885–891, 1992. Podmore, R. et al., Automatic generation control of jointly-owned generating units, IEEE Trans. Power Apparatus Syst., PAS-98 (I), 207–218, January/February, 1979. Post, D., Electric power interchange transaction analysis and selection, Master’s thesis, Iowa State University, Ames, IA, 1994.

15-18

Power Systems

Post, D., S. Coppinger, and G. Sheblé, Application of auctions as a pricing mechanism for the interchange of electric power, IEEE Trans. Power Syst., 10(3), 1580–1584, 1995. Rakic, M.V. and Z.M. Markovic, Short term operation and power exchange planning of hydro-thermal power systems, IEEE Trans. Power Syst., 9(1), 359–365, 1994. Rau, N.S., Certain considerations in the pricing of transmission service, IEEE Trans. Power Syst., 4(3), 1133–1139, 1989. Richter, C. and G. Sheblé, Genetic algorithm evolution of utility bidding strategies for the competitive marketplace, 1997 IEEE/PES Summer Meeting, Berlin, Germany, PE-752-PWRS-1-05-1997, IEEE, New York, 1997a. Richter, C. and G. Sheblé, Building fuzzy bidding strategies for the competitive generator, Proceedings of the 1997 North American Power Symposium, Ames, IA, 1997b. Richter, C. and G. Sheblé, Bidding strategies that minimize risk with options and futures contracts, Proceedings of the 1998 American Power Conference, Session 25, Open Access II-Power Marketing, Paper C, Chicago, IL, 1998. Roy, S., Goal-programming approach to optimal price determination for inter-area energy trading, Intl. J. Energ. Res., 17, 847–862, 1993. Rupanagunta, P., M.L. Baughman, and J.W. Jones, Scheduling of cool storage using non-linear programming techniques, IEEE Trans. Power Syst., 10(3), 1279–1285, 1995. Russel, T., Working with an independent grid in the UK—A generator’s view, Proceedings of the 24th Annual North American Power Symposium, Manhattan, KS, pp. 270–275, September 1992. Schweppe, F.C., M.C. Caramanis, R.D. Tabors, and R.E. Bohn, Spot Pricing of Electricity, Kluwer Academic Publishers, Boston, MA, 1988. Seifi, H. and M.S. Sepasian, Electric Power System Planning: Issues, Algorithms and Solutions, Springer, Heidelberg, Germany, 2001. Sheblé, G.B., Electric energy in a fully evolved marketplace, Proceedings of the 26th Annual North American Power Symposium, Manhattan, KS, pp. 81–90, September 1994a. Sheblé, G.B., Simulation of discrete auction systems for power system risk management, Proceedings of the 27th Annual Frontiers of Power Conference, Oklahoma State University, Stillwater, OK, pp. I.1–I.9, 1994b. Sheblé, G.B., Priced based operation in an auction market structure, Presented at the 1996 IEEE Winter Power Meeting, Baltimore, MD, pp. 1–7, 1996. Sheblé, G. and G. Fahd, Unit commitment literature synopsis, IEEE Trans. Power Syst., 9, 128–135, 1994. Sheblé, G. and J. McCalley, Discrete auction systems for power system management, presented at the 1994 National Science Foundation Workshop, Pullman, WA, 1994. Sheblé, G.B., Auction Methods for Restructured Power System Operation, Springer, 1999a. Sheblé, G.B., Computational Auction Mechanisms for Restructured Power Industry Operation, Springer, 1999b. Shirmohammadi, D., P.R. Gribik, T.K. Law, J.H. Malinowski, and R.E. O’Donnell, Evaluation of transmission network capacity use for wheeling transactions, IEEE Trans. Power Syst., 4(4), 1405–1413, 1989. Skeer, J., Highlights of the international energy agency conference on advanced technologies for electric demand-side management, Proceedings of the Advanced Technologies for Electric Demand-Side Management, International Energy Agency, Sorrento, Italy, 1991. Smith, V.L., Electric power deregulation: Background and prospects, Contemporary Policy Issues, 6, 14–24, 1988. Smith, S., Linear programming model for real-time pricing of electric power service, Oper. Res., 41, 470–483, 1993. Sullivan, R.L., Power System Planning, McGraw-Hill, New York, 1977. Svoboda, A. and S. Oren, Integrating price-based resources in short-term scheduling of electric power systems, IEEE Trans. Energy Convers., 9, 760–769, 1994.

Planning Environments

15-19

Tabors, R.D., Transmission system management and pricing: New paradigms and international comparisons, Paper WM110-7 presented at the IEEE/PES Winter Meeting, T-PWRS, February 1994. Vojdani, A., C. Imparto, N. Saini, B. Wollenberg, and H. Happ, Transmission access issues, presented at the 1995 IEEE/PES Winter Meeting, 95 WM 121-4 PWRS, IEEE, New York, 1994. Vojdani, A., C. Imparto, N. Saini, B. Wollenberg, and H. Happ, Transmission access issues, presented at the 1995 IEEE/PES Winter Meeting, 95 WM 121-4 PWRS, IEEE, New York, 1995. Wang, L., Approximate confidence bounds on monte carlo simulation results for energy production, IEEE Trans. Power Syst., 4(1), 69–74, 1989. Wang, C. and J.R. McDonald, Modern Power System Planning, McGraw-Hill, New York, 1994. Wei, D.C. and N. Chen, Air-conditioner direct load control by multi-pass dynamic programming, IEEE Trans. Power Syst., 10(1), 307–313, 1995. Willis, H.L., Spatial Electric Load Forecasting, Marcel Dekker, New York, pp. 14–17, 1996. Winston, W.E., and C.A. Gibson, Geographical load shift and its effect on interchange evaluation, IEEE Trans. Power Syst., 3(3), 865–871, 1988. Wood, A.J. and B.F. Wollenberg, Power Generation, Operation, and Control, 2nd edn., John Wiley & Sons, New York, 1996. Wu, F. and P. Varaiya, Coordinated multi-lateral trades for electric power networks: Theory and implementation, University of California at Berkeley Research Report, June 1995.

16 Short-Term Load and Price Forecasting with Artificial Neural Networks* 16.1 Artificial Neural Networks............................................................ 16-1 Error Back-Propagation Learning Rule • Adaptive Update of the Weights during Online Forecasting

16.2 Short-Term Load Forecasting........................................................ 16-3 ANNSTLF Architecture • Humidity and Wind Speed • Holidays and Special Days • Performance

Alireza Khotanzad

16.3 Short-Term Price Forecasting........................................................16-9

Southern Methodist University

References................................................................................................... 16-11

Architecture of Price Forecaster • Performance

16.1 Artificial Neural Networks Artificial neural networks (ANN) are systems inspired by research into how the brain works. An ANN consists of a collection of arithmetic computing units (nodes or neurons) connected together in a network of interconnected layers. A typical node of an ANN is shown in Figure 16.1. At the input side, there are a number of so-called connections that have a weight of Wij associated with them. The input denoted by Xi gets multiplied by Wij before reaching node j via the respective connection. Inside the neuron, all the individual inputs are first summed up. The summed inputs are passed through a nonlinear singleinput, single-output function “S” to produce the output of the neuron. This output in turn is propagated to other neurons via corresponding connections. While there are a number of different ANN architectures, the most widely used one (especially in practical applications) is the multilayer feed-forward ANN, also known as a multilayer perceptron (MLP), shown in Figure 16.2. An MLP consists of n input nodes, h so called “hidden layer” nodes (since they are not directly accessible from either input or output side), and m output nodes connected in a feed-forward fashion. The input layer nodes are simple data distributors whereas neurons in the hidden and output layers have an S-shaped nonlinear transfer function known as the “sigmoid activation function,” f(z) = 1/1 + e−z where z is the summed inputs. For hidden layer nodes, the output is Hj =

1

1 + exp −

∑

n

Wij Xi

i =1

* This work was supported in part by the Electric Power Research Institute and 1997 Advanced Technology Program of the State of Texas.

16-1

16-2

Power Systems Inputs

Output

Xi

Wij S O

∑

O = S (∑)

FIGURE 16.1 Model of one node of an ANN.

4

5

W14

W24

W15

W34

1

W11

W35

W25

3

2

W12

W21

X1

W13

W22

W31

W23

W32

X2

W33

X3

FIGURE 16.2 An example of an MLP with three input, three hidden, and two output nodes.

where Hj is the output of the jth hidden layer node, j = 1,…, h, and Xi represents the ith input connected to this hidden node via Wij with i = 1,…, n. The output of the kth output node is given by Yk =

1

1 + exp −

∑

h

Wjk H j

j =1

where Y k is the output of the kth output layer node with k = h + 1,…, m, and Wjk representing connection weights from hidden to output layer nodes. One of the main properties of ANNs is the ability to model complex and nonlinear relationships between input and output vectors through a learning process with “examples.” During learning, known input-output examples, called the training set, are applied to the ANN. The ANN learns by adjusting or adapting the connection weights through comparing the output of the ANN to the expected output. Once the ANN is trained, the extracted knowledge from the process resides in the resulting connection weights in a distributed manner. A trained ANN can generalize (i.e., produce the expected output) if the input is not exactly the same as any of those in the training set. This property is ideal for forecasting applications where some historical data exists but the forecast indicators (inputs) may not match up exactly with those in the history.

Short-Term Load and Price Forecasting with Artificial Neural Networks

16-3

16.1.1 Error Back-Propagation Learning Rule The MLP must be trained with historical data to find the appropriate values for Wij and the number of required neurons in the hidden layer. The learning algorithm employed is the well-known error backpropagation (BP) rule (Rumelhart and McClelland, 1986). In BP, learning takes place by adjusting Wij. The output produced by the ANN in response to inputs is repeatedly compared with the correct answer. Each time, the Wij values are adjusted slightly in the direction of the correct answers by back-propagating the error at the output layer through the ANN according to a gradient descent algorithm. To avoid overtraining, the cross-validation method is used. The training set is divided into two sets. For instance, if three years of data is available, it is divided into a two-year and a one-year set. The first set is used to train the MLP and the second set is used to test the trained model after every few hundred passes over the training data. The error on the validation set is examined. Typically this error decreases as the number of passes over the training set is increased until the ANN is overtrained, as signified by a rise in this error. Therefore, the training is stopped when the error on the validation set starts to increase. This procedure yields the appropriate number of epochs over the training set. The entire three years of data is then used to retrain the MLP using this number of epochs. In a forecasting application, the number of input and output nodes is equal to the number of utilized forecast indicators and the number of desired outputs, respectively. However, there is no theoretical approach to calculate the appropriate number of hidden layer nodes. This number is determined using a similar approach for training epochs. By examining the error over a validation set for a varying number of hidden layer nodes, a number yielding the smallest error is selected.

16.1.2 Adaptive Update of the Weights during Online Forecasting A unique aspect of the MLPs used in the forecasting systems described in this section is the adaptive update of the weights during online operation. In a typical usage of an MLP, it is trained with the historical data and the weights of the trained MLP are then treated as fixed parameters. This is an acceptable procedure for many applications. However, if the modeled process is a nonstationary one that can go through rapid changes, e.g., variations of electric load due to weather swings or seasonal changes, a tracking mechanism with sensitivity to the recent trends in the data can aid in producing better results. To address this issue, an adaptive weight adjustment strategy that takes place during online operation is utilized. The MLP is initially trained using the BP algorithm; however, the trained weights are not treated as static parameters. During online operation, these weights are adaptively updated on a sampleby-sample basis. Before forecasting for the next instance, the forecasts of the past few samples are compared to the actual outcome (assuming that actual outcome for previous forecasts have become available) and a small scale error BP operation is performed with this data. This mini-training with the most recent data results in a slight adjustment of the weights and biases them toward the recent trend in data.

16.2 Short-Term Load Forecasting The daily operation and planning activities of an electric utility requires the prediction of the electrical demand of its customers. In general, the required load forecasts can be categorized into short-term, mid-term, and long-term forecasts. The short-term forecasts refer to hourly prediction of the load for a lead time ranging from 1 h to several days out. The mid-term forecasts can either be hourly or peak load forecasts for a forecast horizon of one to several months ahead. Finally, the long-term forecasts refer to forecasts made for one to several years in the future. The quality of short-term hourly load forecasts has a significant impact on the economic operation of the electric utility since many decisions based on these forecasts have significant economic consequences. These decisions include economic scheduling of generating capacity, scheduling of fuel purchases, system security assessment, and planning for energy transactions. The importance of accurate

16-4

Power Systems

load forecasts will increase in the future because of the dramatic changes occurring in the structure of the utility industry due to deregulation and competition. This environment compels the utilities to operate at the highest possible efficiency, which, as indicated above, requires accurate load forecasts. Moreover, the advent of open access to transmission and distribution systems calls for new actions such as posting the available transmission capacity (ATC), which will depend on the load forecasts. In the deregulated environment, utilities are not the only entities that need load forecasts. Power marketers, load aggregators, and independent system operators (ISO) will all need to generate load forecasts as an integral part of their operation. This section describes the third generation of an ANN hourly load forecaster known as Artificial Neural Network Short-Term Load Forecaster (ANNSTLF). ANNSTLF, developed by Southern Methodist University and PRT, Inc. under the sponsorship of the Electric Power Research Institute (EPRI), has received wide acceptance by the electric utility industry and is presently being used by over 40 utilities across the U.S. and Canada. Application of the ANN technology to the load forecasting problem has received much attention in recent years (Dillon et al., 1991; Park et al., 1991; Ho et al., 1992; Lee et al., 1992; Lu et al., 1993; Peng et al., 1993; Papalexopolos et al., 1994; Khotanzad et al., 1995, 1996, 1997, 1998; Mohammed et al., 1995; Bakirtzis et al., 1996). The function learning property of ANNs enables them to model the correlations between the load and such factors as climatic conditions, past usage pattern, the day of the week, and the time of the day, from historical load and weather data. Among the ANN-based load forecasters discussed in published literature, ANNSTLF is the only one that is implemented at several sites and thoroughly tested under various real-world conditions. A noteworthy aspect of ANNSTLF is that a single architecture with the same input-output structure is used for modeling hourly loads of various size utilities in different regions of the country. The only customization required is the determination of some parameters of the ANN models. No other aspects of the models need to be altered.

16.2.1 ANNSTLF Architecture ANNSTLF consists of three modules: two ANN load forecasters and an adaptive combiner (Khotanzad et al., 1998). Both load forecasters receive the same set of inputs and produce a load forecast for the same day, but they utilize different strategies to do so. The function of the combiner module is to mix the two forecasts to generate the final forecast. Both of the ANN load forecasters have the same topology with the following inputs: • 24 hourly loads of the previous day • 24 hourly weather parameters of the previous day (temperatures or effective temperatures, as discussed later) • 24 hourly weather parameters forecasts for the coming day • Day type indices The difference between the two ANNs is in their outputs. The first forecaster is trained to predict the regular (base) load of the next day, i.e., the 24 outputs are the forecasts of the hourly loads of the next day. This ANN will be referred to as the “Regular Load Forecaster (RLF).” On the other hand, the second ANN forecaster predicts the change in hourly load from yesterday to today. This forecaster is named the “Delta Load Forecaster (DLF).” The two ANN forecasters complement each other because the RLF emphasizes regular load patterns whereas the DLF puts stronger emphasis on yesterday’s load. Combining these two separate forecasts results in improved accuracy. This is especially true for cases of sudden load change caused by weather fronts. The RLF has a tendency to respond slowly to rapid changes in load. On the other hand, since the DLF takes yesterday’s load as the basis and predicts the changes in that load, it has a faster response to a changing situation.

Short-Term Load and Price Forecasting with Artificial Neural Networks

Actual load of day k Actual weather parameters of day k

+ 24

RLF

RLS combiner

Forecast weather parameters of day k +1 Day type index Actual load of day k Actual weather parameters of day k Forecast weather parameters of day k +1 Day type index

DLF

16-5

Final load forecast for day k + 1 24 Outputs

24 Outputs

FIGURE 16.3 Block diagram of ANNSTLF.

To take advantage of the complimentary performance of the two modules, their forecasts are adaptively combined using the recursive least squares (RLS) algorithm (Proakis et al., 1992). The final forecast for each hour is obtained by a linear combination of the RLF and DLF forecasts as

Lˆ k+1(i ) = α B (i ) Lˆ

RLF k +1

(i ) + α C (i ) Lˆ DLF k +1 (i ), i = 1, … , 24

The αB (i) and αC (i) coefficients are computed using the RLS algorithm. This algorithm produces coefficients that minimize the weighted sum of squared errors of the past forecasts denoted by J, N

J=

∑β

N −k

[Lk (i ) − Lˆ k (i )]

2

k =1

where Lk(i) is the actual load at hour i, N is the number of previous days for which load forecasts have been made, and β is a weighting factor in the range of 0 < β ≤ 1 whose effect is to de-emphasize (forget) old data. The block diagram of the overall system is shown in Figure 16.3.

16.2.2 Humidity and Wind Speed Although temperature (T) is the primary weather variable affecting the load, other weather parameters, such as relative humidity (H) and wind speed (W), also have a noticeable impact on the load. The effects of these variables are taken into account through transforming the temperature value into an effective temperature, T__eff, using the following transformation: T__eff = T + α * H

T__eff = T −

W * (65° − T ) 100

16-6

Power Systems

16.2.3 Holidays and Special Days Holidays and special days pose a challenge to any load forecasting program since the load of these days can be quite different from a regular workday. The difficulty is the small number of holidays in the historical data compared to the typical days. For instance, there would be three instances of Christmas Day in a training set of 3 years. The unusual behavior of the load for these days cannot be learned adequately by the ANNs since they are not shown many instances of these days. It was observed that in most cases, the profile of the load forecast generated by the ANNs using the concept of designating the holiday as a weekend day, does resemble the actual load. However, there usually is a significant error in predicting the peak load of the day. The ANNSTLF package includes a function that enables the user to reshape the forecast of the entire day if the peak load forecast is changed by the user. Thus, the emphasis is placed on producing a better peak load forecast for holidays and reshaping the entire day’s forecast based on it. The holiday peak forecasting algorithm uses a novel weighted interpolation scheme. This algorithm will be referred to as “Reza algorithm” after the author who developed it (Khotanzad et al., 1998). The general idea behind the Reza algorithm is to first find the “close” holidays to the upcoming one in the historical data. The closeness criterion is the temperature at the peak-load hour. Then, the peak load of the upcoming holiday is computed by a novel weighted interpolation function described in the following. The idea is best illustrated by an example. Let us assume that there are only three holidays in the historical data. The peak loads are first adjusted for any possible load growths. Let (ti, pi) designate the i-th peak-load hour temperature and peak load, respectively. Figure 16.4 shows the plot of pi vs. ti for an example case. Now assume that t h represents the peak-load hour temperature of the upcoming holiday. t h falls in between t1 and t2 with the implication that the corresponding peak load, ph, would possibly lie in the range of [p1, p2] = R1 + R 2. But, at the same time, t h is also between t1 and t3 implying that ph would lie in [p1, p3] = R1. Based on this logic, ph can lie in either R1 or R1 + R 2. However, note that R1 is common in both ranges. The idea is to give twice as much weight to the R1 range for estimating ph since this range appears twice in pair-wise selection of the historical data points. The next step is to estimate ph for each nonoverlapping interval, R1 and R 2, on the y axis, i.e., [p1, p3] and [p3, p2]. For R1 = [p1, p3] interval:

p3 − p1 pˆ h1 = * (th − t1 ) + p1 t3 − t1

Peak load p2 R2

p3

R1 p1

t1

t2 th

t3 Temperature

FIGURE 16.4 Example of peak load vs. temperature at peak load for a three-holiday database.

Short-Term Load and Price Forecasting with Artificial Neural Networks

16-7

For R 2 = [p3, p2] interval:

p2 − p3 pˆ h 2 = * (th − t3 ) + p3 t 2 − t3

If any of the above interpolation results in a value that falls outside the respective range, Ri, the closest pi, i.e., maximum or minimum of the interval, is used instead. The final estimate of ph is a weighted average of pˆh1 and pˆh2 with the weights decided by the number of overlaps that each pair-wise selection of historical datapoints creates. In this case, since R1 is visited twice, it receives a weighting of two whereas the interval R 2 only gets a weighting coefficient of one. pˆ h =

w1 * pˆ h1 + w 2 * pˆ h 2 w1 + w 2

2 pˆ + 1 pˆ = * h1 * h 2 2 +1

16.2.4 Performance The performance of ANNSTLF is tested on real data from ten different utilities in various geographical regions. Information about the general location of these utilities and the length of the testing period are provided in Table 16.1. In all cases, 3 years of historical data is used to train ANNSTLF. Actual weather data is used so that the effect of weather forecast errors do not alter the modeling error. The testing is performed in a blind fashion meaning that the test data is completely independent from the training set and is not shown to the model during its training. One-to-seven-day-ahead forecasts are generated for each test set. To extend the forecast horizon beyond one day ahead, the forecast load of the previous day is used in place of the actual load to obtain the next day’s load forecast. The forecasting results are presented in Table 16.2 in terms of mean absolute percentage error (MAPE) defined as MAPE =

100 N

N

∑ i =1

Actual(i) − Forecast (i) Actual(i)

with N being the number of observations. Note that the average MAPEs over ten utilities as reported in the last row of Table 16.3 indicate that the third-generation engine is quite accurate in forecasting both hourly and peak loads. In the case of hourly load, this average remains below 3% for the entire forecast horizon of 7 days ahead, and for the peak load it reaches 3% on the seventh day. A pictorial example of one-to-seven-day-ahead load forecasts for utility 2 is shown in Figure 16.5. TABLE 16.1 Utility Information for Performance Study Utility 1 2 3 4 5 6 7 8 9 10

No. Days in Testing Period

Weather Variable

Location

141 131 365 365 134 365 365 365 174 275

T T T,H,W T T T T,H T,H T T,W

Canada South Northeast East Coast Midwest West Coast Southwest South North Midwest

16-8

Power Systems TABLE 16.2 Summary of Performance Results in Terms of MAPE

Utility 1 2 3 4 5 6 7 8 9 10 Average

Days-Ahead

MAPE OF

1

2

3

4

5

6

7

All hours Peak All hours Peak All hours Peak All hours Peak All hours Peak All hours Peak All hours Peak All hours Peak All hours Peak All hours Peak All hours Peak

1.91 1.70 2.72 2.64 1.89 1.96 2.02 2.26 1.97 2.03 1.57 1.82 2.29 2.42 2.22 2.38 1.63 1.83 2.32 2.15 2.05 2.12

2.29 2.11 3.44 3.33 2.25 2.26 2.37 2.59 2.38 2.36 1.86 2.25 2.79 2.78 2.91 3.00 2.04 2.25 2.97 2.75 2.53 2.57

2.53 2.39 3.63 3.46 2.38 2.41 2.51 2.69 2.61 2.49 1.99 2.38 2.90 2.90 3.15 3.12 2.20 2.36 3.25 2.93 2.72 2.71

2.71 2.62 3.77 3.37 2.45 2.49 2.58 2.83 2.66 2.37 2.08 2.50 3.00 2.98 3.28 3.29 2.32 2.51 3.38 3.08 2.82 2.80

2.87 2.73 3.79 3.42 2.53 2.60 2.61 2.85 2.65 2.49 2.14 2.61 3.05 3.07 3.39 3.40 2.40 2.54 3.44 3.16 2.89 2.89

3.03 2.94 3.83 3.52 2.58 2.69 2.65 2.93 2.65 2.51 2.17 2.62 3.10 3.17 3.45 3.45 2.41 2.64 3.52 3.27 2.94 2.97

3.15 3.10 3.80 3.40 2.65 2.82 2.69 2.94 2.74 2.55 2.18 2.63 3.18 3.28 3.50 3.52 2.50 2.78 3.56 3.27 2.99 3.03

TABLE 16.3 Training and Test Periods for the Price Forecaster Performance Study Database

Training Period

Test Period

MAE of Day-Ahead Hourly Price Forecasts ($)

CALPX PJM

Apr 23, 1998–Dec 31, 1998 Apr 2, 1997–Dec 31, 1997

Jan 1, 1999–Mar 3, 1999 Jan 2, 1998–Mar 31, 1998

1.73 3.23

3800 3600 3400 3200 3000 2800 2600 2400

Actual Forecast

2200 2000

20

40

60

80 Hour

100

FIGURE 16.5 An example of a one-to-seven-day-ahead load forecast.

120

140

160

Short-Term Load and Price Forecasting with Artificial Neural Networks

16-9

As pointed out earlier, all the weather variables (T or T__eff) used in these studies are the actual data. In online usage of the model, weather forecasts are used. The quality of these weather forecasts vary greatly from one site to another. In our experience, for most cases, the weather forecast errors introduce approximately 1% of additional error for 1–2 days out load forecasts. The increase in the error for longer range forecasts is more due to less accurate weather forecasts for three or more days out.

16.3 Short-Term Price Forecasting Another forecasting function needed in a deregulated and competitive electricity market is prediction of future electricity prices. Such forecasts are needed by a number of entities such as generation and power system operators, wholesale power traders, retail market and risk managers, etc. Accurate price forecasts enable these entities to refine their market decisions and energy transactions leading to significant economic advantages. Both long-term and short-term price forecasts are of importance to the industry. The long-term forecasts are used for decisions on transmission augmentation, generation expansion, and distribution planning whereas the short-term forecasts are needed for daily operations and energy trading decisions. In this work, the emphasis will be on short-term hourly price forecasting with a horizon extending up to the next 24 h. In general, energy prices are tied to a number of parameters such as future demand, weather conditions, available generation, planned outages, system reserves, transmission constraints, market perception, etc. These relationships are nonlinear and complex and conventional modeling techniques cannot capture them accurately. In a similar manner to load forecasting, ANNs could be utilized to “learn” the appropriate relationships. Application of ANN technology to electricity price forecasting is relatively new and there are few published studies on this subject (Szkuta et al., 1999). The adaptive BP MLP forecaster described in the previous section is used here to model the relationship of hourly price to relevant forecast indicators. The system is tested on data from two power pools with good performance.

16.3.1 Architecture of Price Forecaster The price forecaster consists of a single adaptive BP MLP with the following inputs: • • • •

Previous day’s hourly prices Previous day’s hourly loads Next day’s hourly load forecasts Next day’s expected system status for each hour

The expected system status input is an indicator that is used to provide the system with information about unusual operating conditions such as transmission constraints, outages, or other subjective matters. A bi-level indicator is used to represent typical vs. atypical conditions. This input allows the user to account for his intuition about system condition and helps the ANN better interpret sudden jumps in price data that happen due to system constraints. The outputs of the forecaster are the next day’s 24 hourly price forecasts.

16.3.2 Performance The performance of the hourly price forecaster is tested on data collected from two sources, the California Power Exchange (CALPX) and the Pennsylvania-New Jersey-Maryland ISO (PJM). The considered price data are the Unconstrained Market Clearing Price (UMCP) for CALPX, and Market Clearing Price (MCP) for PJM. The average of Locational Marginal Prices (LMP) uses a single MCP for PJM. The training and test periods for each database are listed in Table 16.3. Testing is performed in a blind fashion, meaning that the test data is completely independent from the training set and is not shown to the model during its training. Also, actual load data is used in place of load forecast.

16-10

Power Systems TABLE 16.4 Results of Performance Study for the Test Period Database

MAE of Day-Ahead Hourly Price Forecasts ($)

Sample Mean of Actual Hourly Prices ($)

Sample Standard Deviation of Actual Hourly Prices ($)

CALPX PJM

1.73 3.23

19.98 17.44

5.45 7.67

The day-ahead forecast results are presented in the first column of Table 16.4 in terms of mean absolute error (MAE) expressed in dollars. This measure is defined as MAE =

100 N

N

∑ Actual Price(i) − Forecast Price(i) i =1

with N being the total number of hours in the test period. To put these results in perspective, the sample mean and standard deviation of hourly prices in the test period are also listed in Table 16.4. Note the correspondence between MAE and the standard deviation of data, i.e., the smaller standard deviation results in a lower MAE and vice versa. Figures 16.6 and 16.7 show a representative example of the performance for each of the databases. It can be seen that the forecasts closely follow the actual data. 42

Actual price ANN forecast

Price ($/MW)

37 32 27 22 17 12 1/6/99

1/7/99

1/8/99

1/9/99

1/10/99

1/11/99

1/12/99

FIGURE 16.6 An example of the ANN price forecaster performance for CALPX price data.

Actual price ANN forecast

Price ($/MW)

53 43 33 23 13 3 1/4/98

1/5/98

1/6/98

1/7/98

1/8/98

1/9/98

FIGURE 16.7 An example of the price forecaster performance for PJM price data.

1/10/98

Short-Term Load and Price Forecasting with Artificial Neural Networks

16-11

References Bakirtzis, A.G. et al., A neural network short term load forecasting model for the Greek power system, IEEE Trans. Power Syst., 11, 2, 858–863, May 1996. Dillon, T.S., Sestito, S., and Leung, S., Short term load forecasting using an adaptive neural network, Int. J. Electr. Power Energy Syst., 13, 4, 186–192, Aug 1991. Ho, K., Hsu, Y., and Yang, C., Short term load forecasting using a multi-layer neural network with an adaptive learning algorithm, IEEE Trans. Power Syst., 7, 1, 141–149, Feb 1992. Khotanzad, A., Afkhami-Rohani, R., Lu, T.L., Davis, M.H., Abaye, A., and Maratukulam, D.J., ANNSTLF—A neural network-based electric load forecasting system, IEEE Trans. Neural Netw., 8, 4, 835–846, July 1997. Khotanzad, A., Afkhami-Rohani, R., and Maratukulam, D., ANNSTLF—Artificial neural network shortterm load forecaster-generation three, IEEE Trans. Power Syst., 13, 4, 1413–1422, Nov 1998. Khotanzad, A., Davis, M.H., Abaye, A., and Maratukulam, D.J., An artificial neural network hourly temperature forecaster with applications in load forecasting, IEEE Trans. Power Syst., 11, 2, 870–876, May 1996. Khotanzad, A., Hwang, R.C., Abaye, A., and Maratukulam, D., An adaptive modular artificial neural network hourly load forecaster and its implementation at electric utilities, IEEE Trans. Power Syst., 10, 3, 1716–1722, Aug 1995. Lee, K.Y., Cha, Y.T., and Park, J.H., Short-term load forecasting using an artificial neural network, IEEE Trans. Power Syst., 7, 1, 124–132, Feb 1992. Lu, C.N., Wu, N.T., and Vemuri, S., Neural network based short term load forecasting, IEEE Trans. Power Syst., 8, 1, 336–342, Feb 1993. Mohammed, O. et al., Practical experiences with an adaptive neural network short-term load forecasting system, IEEE Trans. Power Syst., 10, 1, 254–265, Feb 1995. Papalexopolos, A.D., Hao, S., and Peng, T.M., An implementation of a neural network based load forecasting model for the EMS, IEEE Trans. Power Syst., 9, 4, 1956–1962, Nov 1994. Park, D.C., El-Sharkawi, M.A., Marks, R.J., Atlas, L.E., and Damborg, M.J., Electric load forecasting using an artificial neural network, IEEE Trans. Power Syst., 442–449, May 1991. Peng, T.M., Hubele, N.F., and Karady, G.G., Advancement in the application of neural networks for shortterm load forecasting, IEEE Trans. Power Syst., 8, 3, 1195–1202, Feb 1993. Proakis, J.G., Rader, C.M., Ling, F., and Nikias, C.L., Advanced Digital Signal Processing, Macmillan Publishing Company, New York, 1992, pp. 351–358. Rumelhart, D.E. and McClelland, J.L., Parallel Distributed Processing, Vol. 1, MIT Press, Cambridge, MA, 1986. Szkuta, B.R., Sanabria, L.A., and Dillon, T.S., Electricity price short-term forecasting using artificial neural networks, IEEE Trans. Power Syst., 14, 3, 851–857, Aug 1999.

17 Transmission Plan Evaluation: Assessment of System Reliability 17.1 Bulk Power System Reliability and Supply Point Reliability......... 17-1 Bulk Transmission Systems Reliability Is Evaluated Using Deterministic Reliability Criteria • Supply Point Reliability Is Evaluated Using Either Deterministic or Probabilistic Reliability Criteria

17.2 Methods for Assessing Supply Point Reliability......................... 17-4 Reliability Measures—Reliability Indices • System Indices • Cost of Interruptions to Consumers • Outage Models

N. Dag Reppen Niskayuna Power Consultants, LLC

James W. Feltes Siemens Power Technologies International

17.3 Probabilistic Reliability Assessment Methods............................ 17-8 Contingency Enumeration Approach • Comparison of Contingency Enumeration and Monte Carlo Simulation

17.4 Application Examples................................................................... 17-10 Calculation of the Reliability of Electric Power Supply to a Major Industrial Complex • Local Area Reliability

References................................................................................................... 17-14

17.1 Bulk Power System Reliability and Supply Point Reliability Transmission systems must meet performance standards and criteria that ensure an acceptable level of quality of electric service. Service quality means continuity of supply and constancy of voltage waveform and power system frequency. Frequency is typically not an issue in large interconnected systems with adequate generation reserves. Similarly, voltage quality at the consumer connection is typically addressed at the distribution level and not by reinforcing the transmission system. This leaves continuity of power supply as the main criterion for acceptable transmission system performance. Requirements for continuity of supply are traditionally referred to as power system reliability. Reliability criteria for transmission systems must address both local interruptions of power supply at points in the network as well as widespread interruptions affecting population centers or entire regions. Local and widespread interruptions are typically caused by different types of events and require different evaluation approaches. Additional transmission facilities will virtually always increase reliability, but this remedy is constrained by the cost of new facilities and environmental impacts of new construction. Reliability objectives, therefore, must be defined explicitly or implicitly in terms of the value of reliable power supply to the consumer and to society at large. Reflecting the different concerns of local interruptions and 17-1

17-2

Power Systems

widespread interruptions, reliability objectives are different for the bulk transmission system than for the local area transmission or subtransmission systems supplying electric power to electric distribution systems. These two aspects of power system reliability will be referred to as bulk power system reliability (Endrenyi et al., 1982, Parts 1 and 2) and supply point reliability.

17.1.1 Bulk Transmission Systems Reliability Is Evaluated Using Deterministic Reliability Criteria A distinguishing characteristic of bulk transmission systems is that severe disturbances arising in them can have widespread impact. Major failures of bulk transmission systems have resulted in interruption of thousands of MW of load and interruption of service to millions of customers. Three important characteristics of reliable bulk transmission system performance are

1. Low risk of widespread shutdown of the bulk transmission system 2. Confinement of the extent of bulk transmission system shutdown when it occurs 3. Rapid restoration of operation following shutdown of the bulk transmission system

Most interconnected systems have reliability criteria and design standards that explicitly aim at limiting the risk of widespread shutdowns or blackouts. Such criteria may call for transmission reinforcements or limitations of power transfers across the system. The two other characteristics are addressed by sharpening operating command and control functions and improving control and communication facilities. Therefore, transmission system plans are typically evaluated with respect to reliability criteria that are aimed at limiting the risk of system shutdowns. The U.S. National Electric Reliability Council (NERC), formed in response to the 1965 Northeast blackout, has developed basic design criteria aimed at reducing the risk of “instability and uncontrolled cascading” that may lead to system blackouts. The various regional reliability councils have interpreted these requirements in various ways and produced additional criteria and guides to address this problem (NERC, 1988). Deterministic criteria for bulk power systems will typically include the following requirements:

1. Test criteria for simulated tests aimed at avoiding overload cascading and instability, including voltage collapse. These test criteria specify in generic form: a. The system conditions to be tested: e.g., peak load conditions, lines or generators assumed out on maintenance, transfer levels b. The type of failure that initiates a disturbance: e.g., type and location of short circuit c. Assumptions to be applied regarding the operation of protection systems and other control systems d. The allowable limits of system response: line and transformer loading limits, high and low voltage limits, and criteria for stable operation The system must be reinforced to meet these criteria. 2. Requirements to test extreme contingencies such as the simultaneous outage of two or more parallel lines or the loss of entire substations. These tests are made to determine and understand the vulnerability of the system to such events. When critical extreme contingencies are identified, steps should be taken to minimize the risk of occurrence of such events. 3. Criteria and guides for protection system design to reduce the risk of critical protection initiated disturbances and for protection misoperation that may aggravate a serious system condition.

Evaluations of the system response to specified severe but rare types of failure events are labeled deterministic. The likelihood of the event specified is not considered, except in a qualitative way when the criteria were created. Since only a small subset of all potentially critical events can be tested, the tests are sometimes referred to as “umbrella” tests. A system that passes these selected tests is believed to have

17-3

Transmission Plan Evaluation: Assessment of System Reliability

a degree of resiliency that will protect it not only for the specific disturbances simulated, but also for a multitude of other disturbances of similar type and severity.

17.1.2 Supply Point Reliability Is Evaluated Using Either Deterministic or Probabilistic Reliability Criteria Reliability objectives at the local area transmission or subtransmission level focus on the reliability of supply to specific supply points as shown in Figure 17.1. Statistically, the reliability of supply may be expressed in terms of the frequency of occurrence of load interruptions, the amount of load interrupted, and the duration of the interruptions. Frequency of interruptions and MWh not served over a period such as a year are commonly used measures for the observed or predicted reliability of power supply to a particular node in a transmission system. Probabilistic reliability methods are required to predict reliability in these terms (Endrenyi, 1978; Billinton and Allan, 1984; Salvaderi et al., 1990). These methods will typically consider more likely events rather than the more extreme and very rare events that can lead to system shutdown. This is justified since system shutdown occurrences are not frequent enough to significantly impact the reliability measures calculated. While it is practical to perform probabilistic calculations to assess supply point reliability, deterministic simulation tests are also commonly used. As a minimum, deterministic criteria call for load flow testing of all single line and single transformer outages. This is referred to as single contingency testing or N − 1 testing. For each of these outages, no line or transformer shall exceed its emergency rating, and no voltage shall violate specified high and low emergency voltage limits. Violation of these criteria calls for system reinforcements. Exceptions are typically made for supply points with low peak demand where it is judged to be too expensive to provide for redundant service. Some utilities use a peak load criterion such as 25 MW, above which redundant transmission connections to a supply point are called for. 0.6

13.5 0.85

0.1 0.3

1.2

1.41 11.0

0.4

9.3

0.8

0.8 7.3

0.31

2.6 Frequency of interruption (times per year) 15.5 Energy not supplied (MWh per year)

FIGURE 17.1 Prediction of supply point reliability.

17-4

Power Systems

17.2 Methods for Assessing Supply Point Reliability Supply point reliability may be assessed in four different ways in order of increasing complexity:

1. Deterministic: System alternatives must meet criteria specifying allowable system response to specified contingencies. 2. Probabilistic—System Trouble: System alternatives must meet criteria specified in terms of probabilistic reliability indices reflecting risk of unacceptable system response. 3. Probabilistic—Consumer Impact: Same as (2), but criteria are specified in terms of consumer impact such as risk of supply interruption or risk of load curtailment. 4. Cost/Benefit Analysis: This approach is based on the concept that the proper level of service reliability should be defined by the balance of incremental worth of service reliability improvement and incremental cost of providing that improvement. The approach is also referred to as “effectiveness analysis” or “value-based” reliability assessment.

The limitation of the deterministic approach (1) is that it considers only the initial system problems for a few contingencies. These contingencies have typically been selected by committee based on a mixture of judgment, tradition, and experience. If the selected contingencies do not cover all important reliability concerns, the resulting system may be unreliable. If the selected contingencies put undue emphasis on severe but rare events, an unnecessarily expensive system alternative may be selected. The probabilistic approach (2) aims at eliminating the dependency on judgment in the selection of contingencies by attempting to look at all significant contingencies. In addition, it weighs the importance of the results for each contingency according to the severity of the system problems caused by each contingency and the frequency of occurrence of each contingency. Approach (3) looks deeper into the problem, in that it is concerned with the impact on the consumer. However, the criteria used to define an acceptable level of reliability are still judgmental. For example, how many interruptions per year would be acceptable or what percentage of total MWh demand is it acceptable to interrupt or curtail? In the cost/benefit approach (4), the criterion for acceptable reliability is implicit in the methodology used.

17.2.1 Reliability Measures—Reliability Indices Reliability can be measured by the frequency of events having unacceptable impacts on the system or on the consumer, and by the severity and duration of the unacceptable impacts. Thus, there are three fundamental components of reliability measures:

1. Frequency of unacceptable events 2. Duration of unacceptable events 3. Severity of unacceptable events

From these, other measures, such as probability of unacceptable events, can be derived. An expectation index, such as the loss of load expectation (LOLE) index commonly used to measure the reliability of a generating system is, in its nature, a probability measure. While probability measures have proved useful in generation reliability assessment, they may not be as meaningful in assessing the reliability of a transmission system or a combined generation/transmission system. It is, for example, important to differentiate between 100 events which last 1 s and 1 event which lasts 100 s. Since probability measures cannot provide such differentiation, it is often necessary to apply frequency and duration measures when assessing the reliability of transmission systems. Probabilistic reliability measures or indices can express the reliability improvements of added resources and reinforcements quantitatively. However, several indices are required to capture various reliability aspects. There are two major types of indices: system indices and consumer or load indices

Transmission Plan Evaluation: Assessment of System Reliability

17-5

(Guertin et al., 1978; Fong et al., 1989). The former concerns itself with system performance and system effects, the latter with the impact on the consumer. The reliability cost measure used in cost/benefit analysis may be classified as a consumer index.

17.2.2 System Indices Indices suitable for transmission system reliability evaluation may be divided into system problem indices and load curtailment indices. System problem indices measure frequency, duration, probability, and severity of system problems. Some examples: • Frequency of circuit overloads (overloads/year) • Average duration of circuit overloads (hours) • Probability of circuit overloads Load curtailment indices measure severity in terms of load interrupted or curtailed. The salient characteristic of these indices is that the severity of any event, regardless of the system problems resulting from the event, is expressed in terms of load curtailment. From the three fundamental reliability measures (frequency, duration, and load curtailment), a series of derived reliability indices may be defined as illustrated by the following examples. Basic Annual Indices • Frequency of load curtailment F = • Hours of load curtailment D =

i

i

i

i

i

i

i

i

)

−1

)

−1

i

i

−1

∑ FD (h year

∑ FC (MW year ) Energy curtailed E = ∑ FD C (MWh year

• Power curtailed C = •

∑ F (year

i

i

−1

)

where Fi = Frequency of event i (year−1) Di = Duration of event i (h) Ci = MW load curtailed for event i (MW) i = All events for which Ci > 0 Energy curtailment (E), expressed in MWh not served, is often referred to as Energy Not Served (ENS), Expected Energy Not Served (EENS), or Expected Unserved Energy (EUE). Load curtailment indices are sometimes normalized to system size. Two commonly used indices are • Power interruption index CN = C/CMX (year−1) • Energy curtailment index EN = E/CMX (h year−1) where CMX = peak load for system, area, or bus. EN × 60 is referred to as system minutes, the equivalent number of minutes per year of total system shutdown during peak load conditions.

17.2.3 Cost of Interruptions to Consumers The fact that a sudden interruption of very short duration can have a significant impact and that an outage of 4 h may have a significantly more severe impact than two outages of 2 h each, illustrates the limitations of simple aggregated reliability measures such as MWh not served. This is an important

17-6

Power Systems 100 Residential Commercial

Interruption cost $/KW

Industrial 10

1

0.1

1

10

100

1000

Duration in minutes

FIGURE 17.2 Illustration of customer damage functions for residential, commercial, and industrial load for process-oriented industrial load. The cost of very short duration outages may be much higher than shown here.

limitation since the various transmission reinforcement options considered may have dramatically different impacts as far as interruption durations are concerned. Since it is difficult to use a multiparameter measure when comparing reinforcement alternatives, a single aggregate measure is much preferred as long as it includes the main reliability factors of concern. The concept of cost to consumers of unreliability expressed in dollars per year has emerged as a practical measure of reliability when comparing transmission reinforcement alternatives. As a measure, reliability cost has the additional important advantage that it can be aggregated with installation cost and operating cost to arrive at a minimum “total cost” design in a cost/benefit analysis. Conceptually, the annual reliability cost for a group of customers is the aggregated worth the customers put on avoiding load interruptions. In some cases the costs are tangible, allowing reliable dollar cost estimates; in other instances the impacts of interruptions are intangible and subjective, but still real in the eyes of the consumer. Surveys aimed at estimating what consumers would be willing to pay, either in increased rates or for backup service, have been used in the past to estimate the intangible costs of load interruptions. The results of these investigations may be expressed as Customer Damage Functions (CDF), as illustrated in Figure 17.2 (Mackay and Berk, 1978; Billinton et al., 1983). Customer damage functions can be used to estimate the dollar cost of any particular load interruption given the amount of load lost and the duration of the interruption. If a customer damage function can be assigned for each supply point, then a cost of interrupted load may be determined.

17.2.4 Outage Models Generation and transmission outages may be classified in two categories—forced outages and scheduled outages. While forced outages are beyond the control of the system operators, scheduled outages can usually be postponed if necessary to avoid putting the system in a precarious state. These two outage categories must, therefore, be treated separately (Forrest et al., 1985).

17-7

Transmission Plan Evaluation: Assessment of System Reliability

A

7 B

6

1 C

2 5

3

D

4

= Double circuit tower = Circuit breaker

FIGURE 17.3 Sample system illustrating outage types.

17.2.4.1 Forced Outage Models The variety and characteristics of forced outage events may be illustrated with reference to Figure 17.3. Transmission lines and transformers that can be isolated from the system by the opening of circuit breakers are referred to as “elements.” Three categories of forced outage events are recognized:

1. Single Component Outage Event—The outage event involves only one element. For example, a fault on circuit 1, cleared by circuit breakers in a normal manner, would only affect circuit 1. 2. Common Mode Outage Event—This is a multiple element outage event where a single initiating cause results in multiple element outages where the outages are not consequences of each other. For example, a single lightning stroke taking out both circuits of the double circuit line exiting substation B would be a common mode outage. This event results in the simultaneous outage of circuits 1 and 7. 3. Substation-Related Outage Events—This is a multiple element outage event that depends on the protection system response to a fault on a component in the substation or on an element connected to the substation. Examples of substation-related outage events are: a. Stuck breaker—if the breaker common to circuits 1 and 6 is stuck, a fault on either circuits 1 and 6 would result in both circuits out. b. Tapped circuits—a fault on circuit 2 would result in circuits 2–4 going out together. c. Breaker fault—if there is a fault on the breaker common to circuits 1 and 6, both circuits 1 and 6 would be outraged. d. Bus section fault—a fault on the bus section in substation B would outage circuits 1 and 7.

A common mode outage event may be combined with substation-related outage events. For example, a common mode failure of circuits 1 and 2 would result in an outage event encompassing circuits 1–4. Two or more independent outage events from either of the three outage categories may overlap in time, creating more complex outages. Accurate tools for the prediction of reliability measures include most if not all of these outage types.

17-8

Power Systems

17.3 Probabilistic Reliability Assessment Methods Probabilistic reliability assessment tools falls in one of two categories (Endrenyi et al., 1982a,b):

1. The contingency enumeration method 2. The Monte Carlo method

In general, the contingency enumeration method is capable of looking at severe and rare events such as transmission events in great detail, but cannot practically look at many operating conditions. In contrast, the Monte Carlo methods are capable of looking at operating conditions in great detail (Noferi et al., 1975). However, from a computational standpoint, it is not possible to capture with precision the impact of infrequent but severe transmission contingencies. Thus, the two methods are capturing different aspects of the reliability problem.

17.3.1 Contingency Enumeration Approach The contingency enumeration approach to reliability analysis includes the systematic selection and evaluation of disturbances, the classification of each disturbance according to failure criteria, and the accumulation of reliability indices. Contingency enumeration techniques are structured so as to minimize the number of disturbances that need to be investigated in detail. This is achieved by testing, to the extent possible, only those disturbances that are sufficiently severe to cause trouble and sufficiently frequent to impact the risk indices to be computed. The contingency enumeration approach is structured as shown in Figure 17.4. For a specific pre- disturbance condition, a contingency is selected and tested to determine whether the contingency causes any immediate system problem such as a circuit overload or a bus voltage out of limits. If it does not, a new contingency is selected and tested. The occurrence of a system problem may by itself be logged as a failure. However, in many cases, it will be possible to adjust generation or phase shifters to relieve overloads and to adjust generator voltages or transformer taps to bring bus voltages back within range. It is, therefore, of interest to determine whether it is possible to eliminate a system problem by such corrective actions. A failure is logged when corrective actions, short of curtailing consumer loads, are insufficient to eliminate the system problems. The severity of such system problems may be assessed by computing the amount and location of Select contingency No

System problems? Yes Simulate remedial action

No

System problems? Yes Compute severity Form reliability index

FIGURE 17.4 Contingency enumeration approach.

17-9

Transmission Plan Evaluation: Assessment of System Reliability

load curtailment necessary to eliminate the problem. In this way, it is possible to compute supply point reliability indices that measure the frequency, duration, and amount of expected load curtailment. 17.3.1.1 Monte Carlo Approach Monte Carlo methods (Oliveira et al., 1989) may be sequential or non-sequential. The sequential approach simulates the occurrences of random events through time, recognizing the statistical properties of the various types of events. Typically, the time functions of load and planned generation schedules are established for a period of a year. Starting at the beginning of the year, a sequence of forced shutdown and restoration of transmission and generating equipment is then determined based on random sampling in accordance with the statistical characteristics of the equipment failure processes. The response of the power system during equipment outages is simulated by power flow solutions. Whenever a system condition violating predefined failure criteria is encountered, the occurrence and characteristic of this failure is recorded. At the end of 1 year of simulation, parameters describing the “observed” reliability of the system can be determined. These parameters may include frequency of equipment overload, frequency of voltage violations, MWh not served, average duration and severity of specified types of failures, etc. This process is illustrated in Figure 17.5. One year of simulation constitutes one particular sample scenario governed by the random properties of equipment failure. In order to obtain a measure of the inherent reliability of the system, it

Choose hour Determine load

R R

Determine generators on forced outage

No

Determine weather (fair, bad, or stormy)

R

Determine transmission forced outage

R

System problem Yes Simulate remedial action

No

System problem Yes Compute severity Count/save failure results

No

R

End of year? Yes Repeat annual calculations many times Calculate reliability indices R = Random draw from appropriate probability distribution

FIGURE 17.5 Possible computational sequence for Monte Carlo method.

17-10

Power Systems

is necessary to repeat the simulation over the annual period many times and calculate the reliability measures as the mean of the results obtained over the repeated annual simulations. For reliable systems, several hundred annual simulations may be required to obtain convergence in the reliability measures calculated. When using the sequential approach, it is possible to model time dependencies between key variables. This allows elaborate modeling of energy-limited resources such as hydro plants and pumped hydro. It is also possible to simulate environmental effects such as the occurrence of lightning storms that may impact the failure rate of transmission equipment. A brute-force sequential Monte Carlo simulation would be prohibitively time-consuming when applied to large systems. Practical techniques rely on special sampling techniques and acceptable approximations in power system modeling. If time dependencies are not essential, the non-sequential approach may be used. In this case, hours of simulation may be selected at random rather than in sequence. For a specific hour, a precontingency state is established including bus loads and matching generation dispatch. When it can be used, the nonsequential approach is typically much faster than the sequential approach.

17.3.2 Comparison of Contingency Enumeration and Monte Carlo Simulation From the preceding discussion, it is clear that the Monte Carlo method differs from the contingency enumeration method in the way power system states including load, generation dispatch, and component outages are selected. The actual network solution and corrective action models used may be the same or similar for both methods. The major advantage of the Monte Carlo method is the ease with which comprehensive statistics of power system states can be included. This makes the method suitable for computing period reliability indices such as annual indices (CIGRE, 1992). The Monte Carlo method may not be suitable for estimating the probability or frequency of occurrence of infrequent events. The contingency enumeration method may, therefore, be a more practical approach in system design. In comparing system alternatives to strengthen a local area, the Contingency Enumeration approach will provide consistent and real differences in reliability indices computed for specific situations. Unless the reliability of the alternatives are far apart, it would be very time-consuming and perhaps impractical to obtain acceptable differences in reliability by means of the Monte Carlo method. One way to mitigate this problem is to remove time-consuming calculations from the inner loop of the Monte Carlo calculations. In one approach, which is used to assess the reliability of supply to load centers, the impact of rare transmission failures are obtained from precomputed lookup tables of transmission import limits. Using this approach and various sampling techniques, several thousand years of operation can be simulated in minutes.

17.4 Application Examples The techniques described above are presently used for transmission planning by major utilities. The following examples illustrate some of these methods. The first example uses contingency enumeration while the second example uses Monte Carlo techniques.

17.4.1 Calculation of the Reliability of Electric Power Supply to a Major Industrial Complex Contingency enumeration techniques were used to assess the reliability of the power supply to a major manufacturing complex (Reppen et al., 1990). In this analysis, the reliability concerns encompassed system events and conditions that are capable of disturbing or shutting down all or portions of the manufacturing processes. The events of concern included initial interruptions, sustained interruptions, overloads, voltage violations, voltage collapse, and overload cascading. The reliability effects of possible

17-11

Transmission Plan Evaluation: Assessment of System Reliability

Energy curtailment (MWh/year)

system reinforcements in the immediate local power supply area and in the main grid supplying this area were evaluated by a comprehensive probabilistic reliability analysis. The characteristics of the power system were radial feeds to the plant with provisions for automatic and manual switchover to alternative supply in case of loss of voltage on primary supply feeders, an extensive local 132 kV system, and a regional 300 and 420 kV transmission system. Contingency enumeration methods allow detailed modeling of the network, including the modeling of automatic responses of the power system to disturbances such as special relaying schemes for line tripping, generation runback, and load transfer. Three typical categories of outages—single element outages, independent overlapping outages of multiple elements, and dependent multiple element outages—were considered in the reliability studies. The term “element” encompasses generating units as well as “transmission elements” such as transmission lines, transformers, capacitor banks, and static var devices. The reliability computations included network analysis of outages, classification of failure events according to type and severity, and calculation of reliability indices. Reliability indices representing the predicted frequency of each of the types of failure events were computed as well as load interruption and energy curtailment indices. The indices computed are referred to as annualized indices reflecting the reliability level that would be experienced if the precontingency condition considered should exist for an entire year. The full analysis included assessment of existing power supply conditions, impact of system reconfiguration on the reliability of supply, reliability effects of system reinforcements, and impact of conditions in the main grid. Here we will concentrate on the reliability effects of system reinforcements. Two reinforcements were analyzed: construction of a new 132 kV line completing a loop at some distance from the plant and construction of a 300 kV ring connecting several of the power supply buses to the plant. Figure 17.6 presents the results of the investigation using the energy curtailment index defined earlier. The energy curtailment index aggregates the expected loss of energy on an annualized basis. The results indicate that reinforcement A (the remote line) has no significant effect on the reliability of power supply to the plant. Reinforcement B (the 300 kV ring) provides a substantial overall improvement, although there is no significant improvement in the energy curtailment index for the winter case. While some of the suggested means of improving reliability could have been predicted prior to the analysis, the relative effectiveness of the various actions would not be apparent without a formal reliability analysis. Performing a reliability analysis of this type gives excellent insights into the dominating failure phenomena that govern system performance. The detailed contingency information available 120 100 80 60 40 20 0

Existing system

A. Remote 135 kV line Winter

B. 300 kV ring

Summer

FIGURE 17.6 Benefit of system reinforcements. Annualized energy curtailment (MWh/year). Sustained interruptions only.

17-12

Power Systems

promotes understanding of the way systems fail, while the reliability indices computed provide the perspective necessary to make appropriate system design decisions.

17.4.2 Local Area Reliability The second example considers the task of improving the reliability of electric power to a local area such as a city, major industrial complex, or other load center (Reppen, 1998). In simple terms, the reliability of the supply is a function of the following parameters: • The load in the area as it fluctuates over time. • The maximum amount of power that can be imported from the main grid. This import limit varies by maintenance and forced outage of transmission and generating equipment. • Maximum available local generation at any particular time as it is affected by generation maintenance and forced outages. At any particular time, load curtailment will occur if the load exceeds the maximum import capability into the area plus the maximum generation available in the area. Therefore, reliability of supply to consumers in the area can be measured in terms of statistics of load curtailment. Popular load curtailment measures include frequency of interruptions and energy not served (MWh/year). In addition, the expected annual customer interruption cost can be predicted using Monte Carlo techniques that simulate system conditions repeatedly over a time period to develop reliability measures by aggregating and averaging the impacts of individual load curtailment events. This allows the use of interruption cost functions (customer damage function) to estimate the expected annual cost of load interruptions. Accepting calculated annual interruption costs as a realistic measure of the economic impact on the consumer, one might declare a system reinforcement alternative to be justified from a reliability standpoint if the reduction in interruption cost is greater than the net cost of investment and operation. While such a criterion may not necessarily be appropriate in all cases, it should provide a relevant benchmark in most environments. Figure 17.7 shows key components of the power supply picture for a small city with a peak load of 210 MW. The city is supplied by two generators totaling 150 MW, and by a 138 kV double circuit transmission line from the main grid. Prime reinforcement options are as follows:

1. Add a new single circuit transmission line as indicated in Figure 17.7. 2. Add one gas turbine generator of size to be determined. 3. Add two identical gas turbine generators of size to be determined.

Key questions of interest are

1. Can the line addition be justified on the basis of savings in customer interruption cost? 2. What generator capacities will produce the same reliability improvements as the line addition? 3. What size generators can be justified on the basis of customer interruption?

Figure 17.8 shows results obtained from the Monte Carlo calculations along with the annual fixed charges for investment cost and net annual operating costs. Significant observations that can be made from Figure 17.8 are • The additions of a new 138 kV line, a 50 MW generator, or two 25 MW generators have approximately the same interruption cost savings. However, the reason for this is that all three alternatives are an effective overkill, reducing the interruption cost to almost 0. If load is anticipated to grow, the line addition will be the better performer at a lower cost. • One 15–25 MW generator would give much improved performance and at a cost which can be justified (marginally) based on reliability worth as expressed by the interruption cost curves. • The line addition is clearly the most cost-effective alternative.

17-13

Transmission Plan Evaluation: Assessment of System Reliability

Main grid

Proposed new line

Double circuit line

Proposed generator

100 MW

Proposed generator

50 MW Load: Annual peak = 210 MW

FIGURE 17.7 Power supply configuration for a small city. 5500

Interruption cost savings ($1000)

5000 4500

Zero interruption cost

4000 3500 Gen. rel + Int. loads

2500

25 MW

Interruptible load

2000 1500

2 × 25 MW

2 × 15 MW

15 MW

Gen. reliability

1000 500 0

50 MW

New line

3000

Grounding Live line maintenance

0

500

1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 Annual cost ($1000)

FIGURE 17.8 Annual savings in cost of interruption vs. annual combined investment and operating costs of transmission and generation reinforcements and short term measures. Measures in the upper left triangle can be justified on account of savings in interruption cost.

The first set of calculations compared the benefit of reinforcement by transmission or generation. While such additions are typically the most powerful reinforcements from the standpoint of improved reliability, they are typically also the most expensive. In addition, when dealing with small local systems, the natural or cost-effective line and generation additions are often more than what’s needed to satisfy reliability needs for the next few years. This is particularly true for low load growth scenarios. Thus, there is a need for less expensive alternatives that typically will have

17-14

Power Systems

smaller incremental reliability benefits than the addition of transmission lines and generators. The results of four such alternatives are shown in Figure 17.8:

1. Improve reliability performance of existing generators 2. Improve grounding of transmission lines 3. Introduce live line maintenance 4. Interruptible load contracts

All of the short term measures except live line maintenance can be justified as the annual interruption cost is greater than the total expense associated with the reliability improvement. Also, live line maintenance and transmission line grounding have too small an impact to be of interest. The most cost effective short term measure comes from improvements in the reliability performance of the 100 MW generator, closely followed by interruptible load contracts. If both of these short term measures are taken, the reliability improvement matches that obtainable from the addition of a 25 MW gas turbine generator and at a much lower cost. This example illustrates how it is possible to use Monte Carlo reliability calculations to predict and compare the benefit-cost trade-off of transmission and generation reinforcements and various short term measures.

References Billinton, R. and Allan, R.N., Power-system reliability in perspective, IEE J. Electron Power, 30, 231–236, March 1984. Billinton, R., Wacker, G., and Wojczynski, E., Comprehensive bibliography on electrical service interruption costs, IEEE Trans. Power Appar. Syst., PAS-102, 6, 1831–1837, June 1983. CIGRE Task Force 38.03.10-1992, Power System Reliability Analysis Volume 2 Composite Power System Reliability Evaluation, 1992. Discussion of Regional Council Planning Reliability Criteria and Assessment Procedures, a reference document by the North American Electric Reliability Council, June 1988. Endrenyi, J., Reliability Modeling in Electric Power Systems, John Wiley & Sons, Ltd., New York, 1978. Endrenyi, J., Albrecht, P.F., Billinton, R., Marks, G.E., Reppen, N.D., and Salvadori, L., Bulk power system reliability assessment—Why and how? Part 1: Why? IEEE Paper 82WM 147-7, presented at the Winter Power Meeting, New York, February 1–5, 1982a. Endrenyi, J., Albrecht, P.F., Billinton, R., Marks, G.E., Reppen, N.D., and Salvadori, L., Bulk Power system reliability assessment—Why and how? Part 2: How? IEEE Paper 82WM 148-5, presented at the Winter Power Meeting, New York, February 1–5, 1982b. Fong, C.C., Billinton, R., Gunderson, R.O., O’Neill, P.M., Raksany, J., Schneider, Jr., A.W., and Silverstein, B., Bulk system reliability—Measurement and indices, IEEE Trans. Power Appar. Syst., 4, 3, 829–835, August 1989. Forrest, D.W., Albrecht, P.F., Allan, R.N., Bhavaraju, M.P., Billinton, R., Landgren, G.L., McCory, M.F., and Reppen, N.D., Proposed terms for reporting and analyzing outages of electrical transmission and distribution facilities, IEEE Trans. Power Appar. Syst., PAS-104, 2, 337–348, February 1985. Guertin, M.B., Albrecht, P.F., Bhavaraju, M.P., Billinton, R., Jorgensen, G.E., Karas, A.N., Masters, W.E., Patton, A.D., Reppen, N.D., and Spence, R.P., Reliability indices for use in bulk power supply adequacy evaluation, IEEE Trans. Power Appar. Syst., PAS-97, 4, 1097–1103, July/August 1978. Mackay, E.M. and Berk, L.H., Costs of Power Interruptions to Industry Survey Results, CIGRE, Paper 3207, August 30–September 7, 1978. Noferi, P.L., Paris, L., and Salvaderi, L., Monte Carlo Methods for power system reliability evaluations in transmission or generation planning, in Proceedings, 1975 Reliability and Maintainability Symposium, Washington, DC, 1975.

Transmission Plan Evaluation: Assessment of System Reliability

17-15

Oliveira, G.C., Pereira, M.V.F., and Cunha, S.H.F., A Technique for Reducing Computational Effort in Monte-Carlo Based Composite Reliability Evaluation, IEEE, WM 174-4 PWRS, 1989. Overview of Planning Reliability Criteria of the Regional Reliability Councils of NERC, A reference document by the North American Electric Reliability Council, 1988. Reppen, N.D., Balancing investments and operating costs with customer interruption costs to give increased reliability, presented at the IEE Colloquium on Tools and Techniques for Dealing with Uncertainty, London, U.K., January 27, 1998. Reppen, N.D., Carlsen, T., Glende, I., Bostad, B., and Lam, B.P., Calculation of the reliability of electric power supply to a major industrial complex, presented at the 10th Power Systems Computation Conference (PSCC), Graz, Austria, August 19–24, 1990. Salvaderi, L., Allan, R., Billinton, R., Endrenyi, J., Mc Gillis, D., Lauby, M., Manning, P., and Ringlee, R., State of the Art of Composite-System Reliability Evaluation, CICRE Session, paper 38-104, Paris, August 26–September 1, 1990.

18 Power System Planning 18.1 Planning Entities............................................................................. 18-2 18.2 Arenas................................................................................................ 18-2 18.3 The Planning Problem.................................................................... 18-3 Options • Uncertainties • Attributes

18.4 Planning Processes..........................................................................18-8

Hyde M. Merrill Merrill Energy, LLC

Setting Standards or Criteria • Assessment • Generation Planning • Transmission Planning • Least-Cost Planning • Making Choices

References................................................................................................... 18-12

Power system planning is the recurring process of studying and determining what facilities and procedures should be provided to satisfy and promote appropriate future demands for electricity. The electric power system as planned should meet or balance societal goals. These include availability of electricity to all potential users at the lowest possible cost, minimum environmental damage, high levels of safety and reliability, etc. Plans should be technically and financially feasible. Plans also should achieve the objectives of the entity doing the planning, including minimizing risk. The electric power system is a force-at-a-distance energy-conversion system. It consists of three principal elements: • Current- and voltage-producing, transmitting, and consuming hardware • Control and protective devices • Planning, operating, commercial, and regulatory practices and procedures These definitions are very different from would have appeared on these pages 25 years ago. They no doubt will seem quaint 25 years hence. At this writing, the electric power industry worldwide is experiencing its most dramatic changes in two generations. These changes affect planning, but this section is intended as a practical exposition, not as a history lesson or a prophesy. We therefore will focus on how planning is or should be done today, avoiding flights of fantasy into the past or future (Sullivan, 1977; Kahn, 1988; Ringlee, 1989; Stoll et al., 1989). Planning considers: • Options • Uncertainties • Attributes Options are the choices available to the planner. Uncertainties are parameters whose values are not known precisely or cannot be forecast without error. Attributes are measures of “goodness.” Stakeholder objectives are expressed in terms of attributes. Physical, economic, and institutional realities determine how different options and uncertainties affect the attributes. The planning problem is to identify and choose among options, in the presence of uncertainties, so as to maximize or minimize (as the case may be) the attributes. 18-1

18-2

Power Systems

18.1 Planning Entities Planners generally are trained as engineers, economists, civil servants, businessmen, or mathematicians. They do power system planning for the following entities: • Vertically integrated utilities owning generation, transmission, and distribution systems. • Transmission companies, independent system operators (ISO), and regional transmission organizations (RTO). Transmission companies own transmission assets; the latter do not, but may have some responsibility for their planning. • Pools or combinations of vertically integrated utilities. Other organizations do planning studies and higher-level power sector planning. A step removed from the operation and management of the power system, their interest is in seeing that it meets society’s goals: • Various levels of government • International development banks Still other organizations do power system planning studies, but without system responsibility. They wish to understand the market for various services and how they might compete in it—its economics and technical requirements for entry. • Independent power producers (IPP) or nonutility generators (NUG). These include qualifying facilities (QF as defined by the U.S. Public Utilities Regulatory Policy Act of 1978) and exempt wholesale generators (EWG as defined by the U.S. Energy Policy Act of 1992). These are subject to less stringent regulation than are utilities. They neither enjoy monopoly protection nor have an obligation to provide electricity at cost-based tariffs. • Large industrial users. • Commercial middlemen who buy and sell electrical energy services. • Investors. All of these are supported by independent purveyors of planning information. Consultants with specialized analytic skills also do planning studies.

18.2 Arenas Planning is done in several arenas, distinguished by the planning horizon and by the types of options under consideration. These arenas include • Long-term vs. short-term planning. Economists distinguish these by whether capital investment options are considered. For engineers, long-term planning has a distant horizon (perhaps 30 years for generation and half of that for transmission). Short-term planning considers about 5 years. Operations planning is for as short as a few hours and is not treated here. • Generation vs. transmission vs. least-cost planning. Generation and transmission planning focus on supply options. Least-cost planning includes demand-side options for limiting or shaping load. • Products and services. Some entities provide power (kW) and energy (kWh). Others plan the transmission system. Others provide for auxiliary services (voltage and power control, electrical reserves, etc.). Still others plan for diversified services like conservation and load management. Other arenas require engineering and economic skills, but are within the purview of a book on business or policy rather than an engineering handbook. • Competitive markets. Strategic planning is particularly concerned with financial and business plans in competitive markets. • Sector evolution. Defining the form of the future power sector, including the relationships between competitive forces, regulation, and the broadest social objectives, is a particularly vital planning function.

Power System Planning

18-3

18.3 The Planning Problem 18.3.1 Options Power generation, transformer, transmission system, substation, protection, and operation and control options are discussed in other chapters of this handbook. Other options are discussed below. 18.3.1.1 Planning and Operating Standards or Criteria Planning and operating criteria have a dual nature: they are both attributes and options. Here we will emphasize the fact that they are options, subject to change. Though they have no intrinsic value, standards or criteria are important for several reasons. Their consistent application allows independent systems to interconnect electrically in symmetrical relationships that benefit all. Criteria can also eliminate the need for planners to ask constantly, “How much reliability, controllability, etc. do I need to provide?” Criteria include • Maximum acceptable loss-of-load probability (LOLP) or expected unserved demand, minimum required reserve margins, and similar generation planning standards • What constitutes a single contingency (transmission systems are often designed to withstand “any” single contingency) and whether particular single contingencies are excluded because they are unlikely or expensive to forestall • Permissible operating ranges (voltages, power flows, frequency, etc.) in the normal or preventative state, the emergency state, and the restorative state • How criteria are to be measured or applied Most power systems in industrialized nations are designed and operated so that

1. With all elements in service, power flows, voltages, and other parameters are within normal ranges of the equipment 2. The system remains stable after any single contingency 3. Power flows, voltages, and other operating parameters are within emergency ranges following any single contingency

For financial and economic reasons, developing countries choose weaker criteria. 18.3.1.2 Demand Management Demand-side planning often is tied to generation planning because it affects the power and energy that the power plants will need to provide. There is no perfect classification scheme for demand-side options. Some overlapping classifications are • Indirect load control vs. direct load control by the bulk system operator • Power (kW) or energy (kWh) modification or both • Type of end-use targeted Table 18.1 shows the type of load under direct utility control in the U.S. early in the 1980s, when enthusiasm for demand-side options was especially high. One of the most effective examples of load control was reported by a German utility. Typical off-peak winter demand was less than 70% of the peak for the same day. An indirect program promoted storage space heaters that use electricity at night, when demand is low, to heat ceramic bricks. During the day, air forced among the bricks transfers the heat to the living space. Within 5 years the program was so popular that direct control was added to avoid creating nighttime peaks. The winter daily load shape became practically flat. 18.3.1.3 Market and Strategic Options Market and strategic options are also important. These range from buying a block of power from a neighboring utility to commodity trading in electricity futures to mergers, divestitures, and acquisitions.

18-4

Power Systems TABLE 18.1 Appliances and Sectors under Direct Utility Control, United States—1983 Appliance or Sector Electric water heaters Air conditioners Irrigation pumps Space heating Swimming pool pumps Other Total Residential Commercial Industrial Agricultural

Number Controlled

Percent of Total Controlled

648,437 515,252 14,261 50,238 258,993 13,710

43 34 1 3 17 1

1,500,891

100

1,456,212 29,830 588 14,261

97 2 – 1

Source: U.S. Congress, New Electric Power Technologies: Problems and Prospects for the 1990s, OTA-E-246, Office of Technology Assessment, Washington, DC, July 1985.

18.3.2 Uncertainties Uncertainty can seldom be eliminated. Planning and forecasting are linked so that even if the forecasts are wrong, the plans are right (Bjorklund, 1987). 18.3.2.1 Models of Uncertainty (Schweppe, 1973) Probabilistic models, where different outcomes are associated with different probabilities, are valid if the probability structure is known. The events involved must occur often enough for the law of large numbers to apply, or else the probabilities will have little relationship to the frequencies of the outcomes. Generation planners have excellent probabilistic reliability models. (Generation and transmission reliability evaluation are treated in more detail in a separate section.) Unknown-but-bounded (set theoretic) models are used when one or both of the conditions above are not met. For instance, transmission planners design to withstand any of a set of single contingencies, usually without measuring them probabilistically. 18.3.2.2 Demand Growth Planners forecast the use of energy (MWh) for a period (e.g., a year) first. They divide this by the hours in the period to calculate average demand, and divide again by the projected load factor (average MW demand/ peak MW demand) to forecast peak demand. Three techniques are used most often to forecast energy. Extrapolation—Exponential growth (e.g., 4% per year) appears as a straight line on semi-log paper. Planners plot past loads on semi-log paper and use a straight edge to extrapolate to the future. Econometric models—Econometric models quantify relationships between such parameters as economic activity and population and use of electricity. The simplest models are linear or log-linear:

Di = f (P, GDP, etc.) = k1Di −1 + k 2 Pi + k 3GDPi +

(18.1)

where Di is the demand or log(demand) in period i Pi is the population or log(population) in period i GDPi is the gross domestic product or log(gross domestic product) or some measure of local economic activity k1, k 2, k 3, etc. are coefficients

18-5

Power System Planning

Econometric models are developed in a trial-and-error process. Variants of Equation 18.1 are hypothesized and least squares (regression) analysis is used to find values of coefficients that make Equation 18.1 fit historical data. Econometric models are used by first forecasting population, economic activity, etc. and from them calculating future energy demand using Equation 18.1. End-use models—First, the number of households is forecast. Then the per-household penetration of various appliances is projected. The average kWh used by each appliance is estimated and is multiplied by the two previous numbers. The results are summed over all types of appliances. Performance—Extrapolation became suspect after U.S. load forecasts in the 1970s were consistently too high. Econometric modeling is more work but is more satisfying. End-use modeling requires considerable effort but gives the most accurate forecasts of residential load. Real drivers—One fundamental driver for per-capita load growth is the replacement by electricity of other forms of energy use. The second is the creation of new uses of energy that are uniquely satisfied by electricity. During the decades when U.S. electric demand grew at over 7% per year, the demand for all forms of energy (of which electricity is a part) grew at about 2% per year. This obviously could not continue: the two cannot cross. The growth of electricity demand began to drop off about 1955, declining noticeably in the 1970s and thereafter. The drop in load growth was attributed to the oil crises of the 1970s. Post1973 conservation played a part, but by then electricity had captured about all the market share it was going to get by replacement and creation of new demands for energy. In developing countries, both fundamental drivers are limited by the ability of the electric companies to finance the necessary generation and distribution infrastructure, which is very expensive. Demand is also limited by their ability to generate. In industrialized countries, availability of capital and power plant performance are not constraining. In all countries, elasticity reduces demand if electricity is costly. This effect is much stronger in countries with low per-capita income and for energy-intensive industrial load. 18.3.2.3 Fuel and Water In the near term, strikes, weather, and natural disasters can interrupt production or delivery. Fuel inventories and the ability to redispatch provide good hedges. In the intermediate term, government action can make fuel available or unavailable. For instance, in 1978, the U.S. Congress forbade burning natural gas by utilities, perceiving that there was a shortage. The shortage became a glut once the U.S. natural gas market was deregulated. In the long term, any single source of fuel is finite and will run out. British coal, which had fueled the industrial revolution, was shut down in the 1990s because it had been worked out. The more important fuel uncertainties, however, are in price. For instance, Figure 18.1 shows that the price of crude oil doubled in 1974 and again in 1979. Recognizing the high variability, in 1983 the U.S. Department of Energy forecast a fuzzy band instead of a single trajectory (U.S. Dept of Energy, 1983a). $120

1982 $/BBL

$100 $80 $60 $40 $20 $0 1970

1980

FIGURE 18.1 World oil price projections, 1983.

1990

2000

2010

18-6

Power Systems

It is interesting that within a year of the publication of this projection, the price of oil had dropped below the low limit of the band, and it has remained there until this writing. Planners must consider extreme possibilities for all uncertainties. Brazil, Norway, the Pacific Northwest, Quebec, and a number of developing countries are highly dependent on hydropower. Systems usually are planned and operated so that there will not be a shortfall unless one of the worst hydrological years in recorded history recurs. 18.3.2.4 Construction Three major construction uncertainties are: How long will it take to build? How much will it cost? Will the project be completed? A World Bank study of 41 hydro projects revealed that 37% experienced a schedule slip of 30% or more, including 17% with schedule slips of 60%–100% (Crousillat, 1989). Figure 18.2 shows the range of actual vs. budgeted cost for a number of World Bank-financed projects. The distribution is not symmetrical—overruns are much more frequent than under-budget projects. Some of the worst cases in Figure 18.2 data occurred during periods of unexpected high inflation (Crousillat, 1989). A 1983 report projected that the cost of some 40 U.S. nuclear plants scheduled for completion by 1990 would be close to normally distributed, with the least expensive costing a bit under $2000/kW and the most expensive three times higher, at $6000/kW (U.S. Dept of Energy, 1983b). Possibly the most expensive nuclear plant ever built, the Shoreham Plant on Long Island, was completed at a cost of some U.S. $16 billion. It was shut down by the state before producing a single kWh of commercial energy. 18.3.2.5 Technology New technologies are generally less certain than mature technologies in their cost, construction time, and performance. Even mature technologies may have important uncertainties. For example, transmission transfer capability is an important measure of transmission system capability. It is usually expressed as a single number, but it is actually a time-varying random variable. 18.3.2.6 Demand Management Demand management programs are risky, in part because of uncertainty in the public’s response to them. The two major uncertainties are • What fraction of eligible customers will respond to a particular program? • How much will the average customer change his use of electricity? 35% 30%

Projects

25% 20% 15% 10% 5% 0% –100%

–50%

0%

50%

100%

150%

Cost overrun or underrun

FIGURE 18.2 Budget vs. actual costs, power projects in developing countries.

200%

Power System Planning

18-7

These uncertainties are affected strongly by the design of the program, the incentives offered, how it is marketed, etc. Carrying out a carefully designed pilot program can reduce the uncertainty. The pilot program should be done in the region of the future commercial program. 18.3.2.7 Markets and Capital Recovery For many years, vertically integrated utilities were guaranteed the recovery of all costs, including capital invested, plus a modest but sure profit. The customer paid this and absorbed the market uncertainties. At this writing, the regulated monopoly, cost-recovery market is being replaced in many states and countries by a more competitive market. Some market risks are being transferred from customers to utilities, power marketers, generating companies, speculators, and others. This creates new uncertainties. For example, in competitive generation markets it is not known which potential generating units will be built. This affects both transmission and generation planning. 18.3.2.8 Regulation For the foreseeable future, government will play a key role. The uncertainty in what governments will do propagates into uncertainties in profitability of various players, in market entry, in prices, etc. For example, in the 1980s, U.S. state and federal governments encouraged utilities to implement demand-side programs. Program costs, and in some cases costs of foregone sales, were recovered through tariffs. The government interest later switched to competitive markets. These markets do not have such a convenient mechanism for encouraging demand-side management. As a result, demandside programs became less attractive. 18.3.2.9 Severe Events High-risk, low-probability events usually are not considered by standard planning practices. For example, transmission planners design so that the system will withstand any single contingency. Planning procedures, criteria, and methods generally ignore several simultaneous or near- simultaneous contingencies. The power system is not designed to withstand them—whether or not it does is happenstance. In January 1998 an ice storm of unprecedented magnitude struck the northeastern United States and Quebec. Ice on transmission lines greatly exceeded design standards. Many towers collapsed. All lines feeding Montreal, and all lines south and east of the city, were on the ground. The government later announced a high-risk, low-probability standard: the system should be designed and operated to prevent loss of more than half of the Montreal load should such an event recur.

18.3.3 Attributes Attributes measure “goodness” in different ways, from different perspectives. Each stakeholder has objectives; they are expressed in terms of attributes. Customers of various kinds (residential, commercial, industrial, etc.): • Cost of electricity • Other costs absorbed by the customer • Quality of service (reliability, voltage control, etc.) Investors in various providers of energy and services • New capital required • Net income, earnings per share, and other measures of income • Cash flow, coverage ratios, and other measures of cash use and replacement

18-8

Power Systems Options Uncertainties

Power system

Attributes

FIGURE 18.3 Options, uncertainties, and attributes.

Employees • • • •

Security Promotion opportunities Salaries Healthiness and safety of working conditions

Taxpayers • Tax revenues • Expenditures from public funds Neighbors (environmentalists, visitors, local inhabitants, competitors, etc.) • • • • • •

Emissions or thermal discharges Community disruption Employment opportunities Rights-of-way and other intrusions Flooding Measures of market power

The list of attributes given above is not complete, and different attributes are important for different studies. Deciding on the planning objectives and the attributes for a given study is an important initial step in power system planning. Some attributes are measured using complex computer models. For others, approximate or ad hoc models may be adequate or may be the best that is available. The planner calculates how the options and the uncertainties (Figure 18.3) affect the attributes. Standards or criteria are surrogates for some attributes that are difficult or impossible to compute.

18.4 Planning Processes 18.4.1 Setting Standards or Criteria Planning objectives often conflict. For example, maximizing reliability and minimizing environmental impacts generally conflict with minimizing costs. Since all attributes cannot be measured in dollars, achieving the right trade-off can be a difficult socio-technico-economic-institutional problem. Doing this every day would burden the planning process. Having standards avoids having to revisit such judgments continually. For instance, once it is decided that (say) 20% generation reserve provides adequate reliability at an acceptable cost, the planner accepts 20% as a standard and designs to meet it. Testing whether a particular plan meets the reserve criterion is easy; the planner can concentrate on other issues. Standards should be examined from time to time. If society becomes poorer or richer, its pocketbook may speak for lower or higher standards of service. Changes in technology may justify a change in standards—for instance, development of better scrubbers may make it reasonable to insist on reduced SO2 or NOx emissions. Increased reliance on electricity may require more reliability: a proposal to shut off the power throughout the United States for 1 min to salute Edison’s death was quashed. Had he died in 1900 instead of 1931, it might have been practical.

Power System Planning

18-9

18.4.2 Assessment 18.4.2.1 Forecasts and Projections Not all uncertainties create risk for every planning study. Those that do for a particular study are identified. Forecasts and projections are developed for these uncertainties. 18.4.2.2 System State The state assessment begins with an evaluation of the technical and economic attributes of the present and future power system. Does it and will it satisfy established technical standards? Is it economical? Does it meet other objectives? Chapter 8 of this handbook, “Power System Analysis and Simulation,” describes the analytical tools available to planners. It also describes how these tools are used. The phenomena analyzed are described in Chapters 10 through 12.

18.4.3 Generation Planning Chapter 2 of this handbook, “Electric Power Generation: Conventional Methods,” describes generation planning options and their characteristics. The generation planner does a preliminary selection from among them, recognizing any special features of his planning problem. Planners measure how the various options would alleviate deficiencies discovered in the assessment step. The effects of various options or combinations of options and the effects of uncertainties on other attributes are also measured. In particular, planners compute reserve margin or other measures of reliability. They simulate the operation of the system to measure operating cost and to determine if the operation is within acceptable ranges of other parameters.

18.4.4 Transmission Planning Traditional transmission options and new technologies are described in Chapters 3, 4, 11, 15, and 16 of this handbook. Like the generation planner, the transmission planner makes a preliminary selection based on the needs and development pattern of his system. For instance, for technical and commercial reasons, a given system will use only a few distinct voltage classes. So a system whose existing transmission consists of 138, 345, and 765 kV equipment will rarely add new circuits at 230 or 400 kV, even though these may be popular elsewhere. Transmission planners then identify a set of specific options and measure how these options in various combinations, along with the important uncertainties, affect the attributes. Load flow, short circuit, and stability analyses are performed to determine if voltages and currents are within acceptable bounds under various system states, and if the system will remain stable for all contingencies. How often and how much the operation of the generation system will be constrained by transmission limitations is an important consideration.

18.4.5 Least-Cost Planning Least-cost planning is also known as integrated resource planning or integrated demand/supply planning. It considers supply-side options (generally generation options) on a level playing field with demandside options (generally conservation, indirect load shifting, or direct load control). These options include incentives to encourage utilities and consumers to change energy consumption patterns. As with generation planning and transmission planning, a preliminary selection weeds out options that are clearly not of interest in a particular area. The least-cost planning process includes computing values of key attributes for various options and uncertainties.

18-10

Power Systems

18.4.6 Making Choices A key question in generation, transmission, and least-cost planning is: How is one plan selected over another? A few distinctive approaches will be described. 18.4.6.1 Minimize Revenue Requirements The planner selects the best option from the ratepayer’s perspective. He selects the plan that will minimize the ratepayer’s cost of electricity while satisfying reliability, environmental, and other criteria. The ratepayer’s cost—an attribute—is the revenue that the utility will have to collect to recover all operating and capital costs and to earn a commission-approved return on unrecovered investor capital:

RR i (O, U) = FC i (O, U) + VC i (O, U)

(18.2)

where RR i(O, U) is the revenue requirements in period i FCi(O, U) is the fixed costs in period i VCi(O, U) is the variable costs in period i O is the selection of the various options U is the realizations or values of the various uncertainties Fixed costs are independent of how much or how little a piece of equipment is used. Depreciation (recovery of investors’ capital), interest on debt, and profit (return on unrecovered capital) are typical fixed costs. Variable costs—fuel cost—for example, are related to how much a piece of equipment is used. These costs include all system costs, not just the cost of the individual option. For instance, old plants may run less when a new plant is built. The variable cost includes fuel cost for all plants. To apply this traditional method, the planner must know his company’s return rate, which is set by the regulator. In a closely related method, the market defines the cost of electricity and Equation 18.2 is solved for the internal rate of return (IRR). The option selected is the one that maximizes the IRR, the investor’s profit. 18.4.6.2 Cost-Benefit Analysis If the benefits exceed the costs, a project is worth doing. Typically, costs are incurred first, and benefits come later. A dollar of benefit later is not worth the same as a dollar of cost today. Present worth analysis is a way to compare dollars at different times. The basic equation is P = S /(1 + i)n

(18.3)

In Equation 18.3, P is the present worth or equivalent value today of an amount S, n years in the future, with i the discount rate or annual cost of capital. Cost-benefit analysis is also used to rank mutually exclusive projects—the one with the highest benefit/cost ratio wins. 18.4.6.3 Multi-Objective Decision Analysis 18.4.6.3.1 Utility Function Methods Table 18.2 compares two options for a new power plant in Utah (Keeney et al., 1981). Which choice is better? The attributes are combined in a utility function of the form:

U(x) = k1Economics + k 2 Environment + + k 7 Feasibility

(18.4)

18-11

Power System Planning TABLE 18.2 Attributes: Nuclear Plant vs. Coal Plant

Economics ($/MWh) Environment (corridor-miles) Public disbenefits ($ × 000,000) Tax revenues ($ × 000,000/year) Health lost (equivalent years) Public attitudes Feasibility

Wellington Coal Plant

Green River Nuclear Plant

60.7 532.6 15.0 3.5 446.7 0.33 60.0

47.4 500.8 22.6 1.0 6.3 −1.0 37.0

Source: Keeney, R.L. et al., Decision Framework for Technology Choice, Report EA-2153, Electric Power Research Institute, Palo Alto, CA, 1981. With permission.

In Equation 18.4, x takes on one of two values, “coal” or “nuclear.” (The actual functional form for a particular study may be more complicated than Equation 18.4.) The coefficients k i reflect the relative importance of each attribute. These coefficients convert the different attributes to a common measure. The choice that minimizes the utility function wins. In this study, for the values of coefficients selected U(coal) was $131.4/MWh; U(nuclear) was $162.9/MWh. This approach has many variants. Uncertainties can be included, making U(x) a random variable. Work has been done to develop methods for determining the decision-maker’s values (the coefficients) and risk tolerance. 18.4.6.3.2 Trade-Off Analysis Trade-off analysis measures each attribute in its natural units, without reducing them all to a common measure, and seeks reasonable compromises at points of saturation or diminishing return. A good compromise will not necessarily optimize any of the attributes, but will come close to optimizing all of them. For example, Figure 18.4 shows 22 plans examined in an energy strategy study. The plans in region A minimize SO2 emissions, but are very costly. The plans in region B are cheap but have high emissions. The plans at the knee of the trade-off curve are at the point of diminishing returns for one attribute against the other. Significant reductions in emissions can be had at little cost by moving from B to the knee. Going beyond the knee toward A will not reduce SO2 much more but will increase the cost significantly. The plans at the knee come close to minimizing both cost and emissions. Trade-off analysis can be done graphically for two-attribute problems. More than two attributes cannot be graphed easily but can be analyzed mathematically (Crousillat et al., 1993).

Present worth of revenue requirements ($ billions)

44 A

43 42 41 40 39

B 0

1000

4000 2000 3000 Average SO2 emissions (g/s)

FIGURE 18.4 Trade-off: cost vs. SO2 emissions.

5000

18-12

Power Systems

18.4.6.4 Risk Risk is the hazard due to uncertainty. Risk is also associated with decisions. Without uncertainties and alternatives, there is no risk. System planning, engineering, and operating procedures have evolved to reduce the risks of widespread or local service interruptions. Another section of this handbook describes methods for modeling and enhancing reliability. Not all risks are included in reliability analysis, however. Much talk about risk is directed to financial risks. Other important risks are not quantified in dollars. One measure of risk is robustness, the likelihood that a particular decision will not be regretted. Exposure is the possible loss (in terms of an attribute) under adverse realizations of uncertainties. Regret is the difference between the value of an attribute for a particular set of decisions and realizations of uncertainties, and the value of the attribute for optimized decisions with perfect foreknowledge of the uncertainties. Planners develop hedges, options that increase robustness or decrease exposure or regret. Building small generating units instead of large ones is an example; an insurance policy is another (De la Torre et al., 1999).

References Bjorklund, G.J., Planning for uncertainty at an electric utility, Public Utilities Fortnightly, Oct. 15, 1987. Crousillat, E., Incorporating Risk and Uncertainty in Power System Planning, I&ED Energy Series paper # 17, The World Bank, Washington, DC, 1989. Crousillat, E.O., Dörfner, P., Alvarado, P., Merrill, H.M., Conflicting objectives and risk in power system planning, IEEE Trans. Power Syst., 8(3), 887–893, Aug. 1993. De la Torre, T., Feltes, J.W., Gómez, T., and Merrill, H.M., Deregulation, privatization, and competition: Transmission planning under uncertainty, IEEE Trans. Power Syst., 14(2), 460–465, May 1999. Kahn, E., Electric Utility Planning & Regulation, American Council for an Energy-Efficient Economy, Washington, D.C., 1988. Keeney, R.L., Beley, J.R., Fleischauer, P., Kirkwood, C.W., Sicherman, A., Decision Framework for Technology Choice, Report EA-2153, Electric Power Research Institute, Palo Alto, CA, 1981. Ringlee, R.J., Ed., Special section on electric utility systems planning, Proc. IEEE, 77(6), June 1989. Schweppe, F.C., Uncertain Dynamic Systems, Prentice-Hall, Englewood Cliffs, NJ, 1973, chap. 3. Stoll, H.G., Garver, L.J. (sic), Jordan, G.A., Price, W.H., Sigley, R.F., Jr., Szczepanski, R.S., Tice, J.B., LeastCost Electric Utility Planning, John Wiley & Sons, New York, 1989. Sullivan, R.L., Power System Planning, McGraw-Hill, New York, 1977. U.S. Dept of Energy, Energy Projections to the Year 2010, Report DOE/PE-0029/2, Office of Policy, Planning and Analysis, Washington, DC, Oct. 1983a. U.S. Dept of Energy, The Future of Electric Power in America: Economic Supply for Economic Growth, Report DOE/PE-0045, Office of Policy, Planning and Analysis, Washington, DC, June 1983b. U.S. Congress, New Electric Power Technologies: Problems and Prospects for the 1990s, OTA-E-246, Office of Technology Assessment, Washington, DC, July 1985.

19 Power System Reliability NERC Regions.................................................................................. 19-2 System Adequacy Assessment....................................................... 19-3 System Security Assessment..........................................................19-4 Probabilistic Security Assessment................................................19-5 Distribution System Reliability.....................................................19-6 Typical Sequence of Events after an Overhead Distribution Fault............................................................................ 19-7 19.7 Distribution Reliability Indices..................................................... 19-7 19.8 Storms and Major Events................................................................19-9 19.9 Component Reliability Data........................................................ 19-10 19.10 Utility Reliability Problems......................................................... 19-11 19.1 19.2 19.3 19.4 19.5 19.6

Underground Cable • Transformer Failures • Lightning • Tree Contact • Birds • Squirrels • Snakes • Insects • Bears, Bison, and Cattle • Mice, Rats, and Gophers • Vandalism

Richard E. Brown Quanta Technology

19.11 Reliability Economics................................................................... 19-14 19.12 Annual Variations in Reliability................................................. 19-15 References................................................................................................... 19-15

The electric power industry began in the late 1800s as a component of the electric lighting industry. At this time, lighting was the only application for electricity, and homes had other methods of illumination if the electricity supply was interrupted. Electricity was essentially a luxury item and reliability was not an issue. As electricity became more common, new applications began to appear. Examples include electric motors, electric heating, irons, and phonographs. People began to grow accustomed to these new electric appliances, and their need for reliable electricity increased. This trend culminated with the invention of the radio. No nonelectrical appliance could perform the same function as a radio. If a person wanted to listen to the airwaves, electricity was required. As radio sales exploded in the 1920s, people found that reliable electricity was a necessity. By the late 1930s, electricity was regarded as a basic utility (Philipson and Willis, 1999). As electric utilities expanded and increased their transmission voltage levels, they found that they could improve reliability by interconnecting their system to neighboring utilities. This allowed connected utilities to “borrow” electricity in case of an emergency. Unfortunately, a problem on one utility’s system could now cause problems to other utilities. This fact was made publicly evident on November 9, 1965. On this day, a major blackout left cities in the northeastern United States and parts of Ontario without power for several hours. Homes and businesses had become so dependent on electricity that this blackout was crippling. Action was needed to help prevent such occurrences from happening in the future.

19-1

19-2

Power Systems

19.1 NERC Regions The North American Electric Reliability Corporation (NERC) was formed in 1968 as a response to the 1965 blackout. By this time, reliability assessment was already a mature field and was being applied to many types of engineered systems (Billinton and Allan, 1988; Ramakumar, 1993). NERC’s mission is to promote the reliability of the North America’s bulk power system (generation and transmission). It reviews past events; monitors compliance with policies, standards, principles, and guides; and assesses future reliability for various growth and operational scenarios. NERC provides planning recommendations and operating guidelines, but has no formal authority over electric utilities. Since most of the transmission infrastructure in the United States and Canada is interconnected, bulk power reliability must look at systems larger than a single utility. The territory covered by NERC is far too large to study and manage as a whole, and is divided into 10 regions. These NERC regions are as follows: East Central Area Reliability Coordination Agreement (ECAR), Electric Reliability Council of Texas (ERCOT), Florida Reliability Coordinating Council (FRCC), Mid-Atlantic Area Council (MAAC), Mid-Atlantic Interconnected Network (MAIN), Mid-Continent Area Power Pool (MAPP), Northeast Power Coordinating Council (NPCC), Southeastern Electric Reliability Council (SERC), Southwest Power Pool (SPP), and the Western Systems Coordinating Council (WSCC). The geographic territories assigned to the 10 NERC regions are shown in Figure 19.1. Even though there are 10 NERC regions, there are only four major transmission grids in the United States and Canada: the area associated with the WSCC, the area associated with the ERCOT, Quebec, and the eastern United States. These are usually referred to as the Western Interconnection, the ERCOT Interconnection, the Quebec Interconnection, and the Eastern Interconnection. Each of these grids is highly interconnected within their boundaries, but only has weak connections to the other grids. The geographic territories associated with these four interconnections are shown in Figure 19.1. NERC looks at two aspects of bulk power system reliability: system adequacy and system security. A system must have enough capacity to supply power to its customers (adequacy), and it must be able to

NPCC MRO

RFC

WECC SPP

SERC ERCOT

FIGURE 19.1 NERC regions.

FRCC

19-3

Power System Reliability

continue supplying power to its customers if some unforeseen event disturbs the system (security). Each of these two aspects of reliability is further discussed in the following.

19.2 System Adequacy Assessment System adequacy is defined as the ability of a system to supply all of the power demanded by its customers (Billinton and Allan, 1988). Three conditions must be met to ensure system adequacy. First, its available generation capacity must be greater than the demanded load plus system losses. Second, it must be able to transport this power to its customers without overloading any equipment. Third, it must serve its loads within acceptable voltage levels. System adequacy assessment is probabilistic in nature (Allan et al., 1994; Schilling et al., 1989). Each generator has a probability of being available, PA, a probability of being available with a reduced capacity, PR, and a probability of being unavailable, PU. This allows the probability of all generator state combinations to be computed. A simple two-generator example is shown in Table 19.1. There are nine possible generator state combinations, and the probability of being in a particular combination is the product of the individual generator state probabilities. In general, if there are n generators and x possible states for each generator, then the number of possible generator state combinations is

Generator state combinations = x n

(19.1)

In addition to generator state combinations, loading behavior must be known. Information is found by looking at historical load bus demand in recent years. For the best accuracy, 8760 h peak demand curves are used for each load bus. These correspond to hourly peak loads for a typical year. To reduce computational and data requirements, it is usually acceptable to reduce each set of 8760-h load curves to three weekly load curves (168 h each). These correspond to typical weekly load patterns for winter conditions, spring/autumn conditions, and summer conditions. Weekly load curves can be scaled up or down to represent temperatures that are above or below normal. Sample weekly load curves for a winter peaking load bus are shown in Figure 19.2. To perform an adequacy assessment, each generation state combination is compared with all hourly loading conditions. For each combination of generation and loading, a power flow is performed. If the available generation cannot supply the loads or if any constraints are violated, the system is inadequate and certain loads must be shed. After all generation/load combinations are examined, the adequacy assessment is complete. An adequacy assessment produces the following information for each load bus: (1) the combinations of generation and loading that result in load interruptions and (2) the probability of being in each TABLE 19.1 Generator State Probabilities Generator State Generator 1

Generator 2

Probability

Available Available Available Reduced Reduced Reduced Unavailable Unavailable Unavailable

Available Reduced Unavailable Available Reduced Unavailable Available Reduced Unavailable

PA1 PA2 PA1 PR2 PA1 PU2 PR1 PA2 PR1 PR2 PR1 PU2 PU1 PA2 PU1 PR2 PU1 PU2

19-4

Power Systems 100

System loading (% of peak)

90 80 70 60 50 40 Winter Spring/fall Summer

30 20 10 0

Sun.

Mon.

Tues.

Wed.

Thurs.

Fri.

Sat.

FIGURE 19.2 Weekly load curves by season.

of these inadequate state combinations. From this information, it is simple to compute the expected number of interruptions for each load bus, the expected number of interruption minutes for each load bus, and the expected amount of unserved energy for each load bus. These load bus results can then be aggregated to produce the following system indices: • LOLE (loss of load expectation)—the expected number of hours per year that the system will have to shed load • EENS (expected energy not served)—the expected number of megawatt hours per year that a system will not be able to supply System adequacy assessment assumes that the transmission system is available. This may not always be the case. A classic example is the 1965 blackout, which was initiated by the unexpected loss of a transmission line. To address such events, system security assessment is required.

19.3 System Security Assessment System security is defined as the ability of a power system to supply all of its loads in the event of one or more contingencies (a contingency is an unexpected event such as a system fault or a component outage). This is divided into two separate areas: static security assessment and dynamic security assessment. Static security assessment determines whether a power system is able to supply peak demand after one or more pieces of equipment (such as a line or a transformer) are disconnected. The system is tested by removing a piece (or multiple pieces) of equipment from the normal power flow model, rerunning the power flow, and determining if all bus voltages are acceptable and all pieces of equipment are loaded below emergency ratings. If an unacceptable voltage or overload violation occurs, load must be shed for this condition and the system is insecure. If removing any single component will not result in the loss of load, the system is N − 1 Secure. If removing any X arbitrary components will not result in the loss of load, the system is N − X Secure. N refers to the number of components on the system and X refers to the number of components that can be safely removed. Performing a static security assessment can be computationally intensive. For example, an N − 2 assessment on a modest system with 5000 components (1500 buses, 500 transformers, and 3000 lines) will require more than 25 million power flows to be performed. For this reason, contingency ranking methods are often used. These methods rank each contingency based on its likelihood of resulting in

19-5

Power System Reliability

Rotor angle difference (deg)

180 160

Fault not cleared

140

Fault cleared at 0.3 s

Critical angle

120 100

Fault cleared at 0.2 s

80

Fault 60 occurs 40 20 0 –0.2 –0.1

0.0

0.1

0.2

0.3 0.4 Time (s)

0.5

0.6

0.7

0.8

FIGURE 19.3 Dynamic security assessment.

load curtailment. Contingencies are examined in order of their contingency ranking, starting with the most severe. If a prespecified number of contingencies are tested and found to be secure, it is assumed that contingencies with less severe rankings are also secure and do not need to be examined. Static security assessment is based on steady state power flow solutions. For each contingency, it assumes that the system protection has properly operated and the system has reached a steady state. In fact, the power system may not actually reach a steady state after it has been disturbed. Checking whether a system will reach a steady state after a fault occurs is referred to as dynamic security assessment (also referred to as transient security assessment). When a fault occurs, the system is less able to transfer power from synchronous generators to synchronous motors. Since the instantaneous power input has not changed, generators will begin to speed up and motors will begin to slow down (analogous to the chain slipping while riding a bicycle). This increases the rotor angle difference between generators and motors. If this rotor angle exceeds a critical value, the system will become unstable and the machines will not be able to regain synchronism. After the protection system clears the fault, the rotor angle difference will still increase since the power transfer limits of the system are still less than the prefault condition. If the fault is cleared quickly enough, this additional increase will not cause the rotor angle difference to exceed the critical angle and the system will return to a synchronous state (ABB Power, 1997). An example of a transient stability test is shown in Figure 19.3. This shows the rotor angle difference between a synchronous generator and a synchronous motor during a fault sequence. When the fault occurs, the rotor angle begins to increase. If the fault is not cleared, the rotor angle quickly exceeds the critical angle. If the fault is cleared at 0.3 s, the rotor angle still increases beyond the critical value. The system is dynamically stable for this fault if it is cleared in 0.2 s. The rotor angle will still increase after the fault occurs but will stabilize below the critical value. A dynamic security assessment will consist of many transient stability tests that span a broad range of loading conditions, fault locations, and fault types. To reduce the number of tests required, contingency rankings (similar to static security assessment) can be used.

19.4 Probabilistic Security Assessment Although the “N − 1 Criterion” remains popular, it has received much criticism since it treats unlikely events with the same importance as more frequent events. Using the N − 1 Criterion, large amounts of money may be spent to reinforce a system against a very rare event. From a reliability perspective, this money might be better spent in other areas such as replacing old equipment, decreasing maintenance intervals, adding automation equipment, adding crews, and so on. To make such value judgments, both the impact of each contingency and its probability of occurrence must be considered (Endrenyi, 1978).

19-6

Power Systems

This is referred to as probabilistic security assessment. To do this type of assessment, each piece of equipment needs at least two fundamental pieces of information: the failure rate of the equipment (usually denoted λ, in failures per year) and the mean time to repair of the equipment (usually denoted MTTR, in hours). Performing a probabilistic security assessment is similar to a performing a standard static security assessment. First, contingencies are ranked and simulated using a power flow. If a contingency results in the loss of load, information about the number and size of interrupted loads, the frequency of the contingency, and the repair time of the contingency is recorded. This allows quantities such as EENS to be easily computed. If contingency i causes kWi amount of kilowatts to be interrupted, then EENS is equal to EENS =

∑ kW λ MTTR i

i

i

(19.2)

i

It is important to note that this is the EENS due to contingencies and is separate from the EENS due to generation unavailability. It is also important to note that this formula assumes that λiMTTRi is small when compared to 1 year. If this is not the case, a component will experience fewer failures per year than its failure rate and the equation must be adjusted accordingly.

19.5 Distribution System Reliability The majority of customer reliability problems stem from distribution systems. For a typical residential customer with 90 min of interrupted power per year, between 70 and 80 min will be attributable to problems occurring on the distribution system that it is connected to (Billinton and Jonnavitihula, 1996). This is largely due to radial nature of most distribution systems, the large number of components involved, the sparsity of protection devices and sectionalizing switches, and the proximity of the distribution system to end-use customers. Since reliability means different things to different people, it is necessary to address the definition of “distribution system reliability” in more detail. In distribution systems, reliability primarily relates to equipment outages and customer interruptions: • Outage—when a piece of equipment is deenergized • Momentary interruption—when a customer is deenergized for less than a few minutes • Sustained interruption—when a customer is deenergized for more than a few minutes Customers do not, in the strictest sense, experience power outages. Customers experience power interruptions. If power is restored within a few minutes, it is considered a momentary interruption. If not, it is considered a sustained interruption. The precise meaning of “a few minutes” varies from utility to utility, but is typically between 1 and 5 min. The IEEE defines a momentary interruption based on 5 min. (Note: some references classify interruptions into four categories rather than two. Instantaneous interruptions last a few seconds, momentary interruptions last a few minutes, temporary interruptions last a few hours, and sustained interruptions last many hours.) On a historical note, momentary interruptions used to be considered a “power quality issue” rather than a “reliability issue.” It is now generally agreed that momentary interruptions are an aspect of reliability since (1) momentary interruptions can cause substantial problems to all types of customers, and (2) many trade-offs must be made between momentary interruptions and sustained interruptions during system planning, operation, and control. It can also be observed that customer voltage sags, typically considered a power quality issue, are slowly becoming a reliability issue for similar reasons.

Power System Reliability

19-7

Distribution system reliability is not dependent solely upon component failure characteristics. It is also dependent upon how the system responds to component failures. To understand this, it is necessary to understand the sequence of events that occurs after a distribution system fault.

19.6 T ypical Sequence of Events after an Overhead Distribution Fault The following is a typical sequence of events that will occur after a fault occurs on a distribution system (Brown, 2003).

1. The fault causes high currents to flow from the source to the fault location. These high currents may result in voltage sags for certain customers. These sags can occur on all feeders that have a common coupling at the distribution substation. 2. An instantaneous relay trips open the feeder circuit breaker at the substation. This causes the entire feeder to be deenergized. A pause allows the air around the fault to deionize, and then a reclosing relay will close the circuit breaker. If no fault current is detected, the fault has cleared itself and all customers on the feeder have experienced a momentary interruption. 3. If the fault persists, time overcurrent protection devices are allowed to clear the fault. If the fault is on a fused lateral, the fuse will blow and customers on the lateral will be interrupted. If the feeder breaker trips again, the reclosing relay will repeat the reclosing process a preset number of times before locking out. After the feeder breaker locks out, all customers on the feeder will be interrupted. Automated line switching and system reconfiguration will occur at this point if these capabilities exist. 4. The electric utility will receive trouble calls from customers with interrupted power. It will dispatch a crew to locate the fault and isolate it by opening up surrounding sectionalizing switches. It may also attempt to reconfigure the distribution system in an attempt to restore power to as many customers as possible while the fault is being repaired. Fault isolation can be very fast if switches are motor operated and remotely controlled, but switching usually takes between 15 and 60 min. 5. The crew repairs the faulted equipment and returns the distribution system to its normal operating state.

As can be seen, a fault on the distribution system will impact many different customers in many different ways. In general, the same fault will result in voltage sags for some customers, momentary interruptions for other customers, and varying lengths of sustained interruptions for other customers, depending on how the system is switched and how long the fault takes to repair. Distribution system reliability assessment methods are able to predict distribution system reliability based on system configuration, system operation, and component reliability data (Brown et al., 1996). This ability is becoming increasingly important as the electric industry becomes more competitive, as regulatory agencies begin to regulate reliability, and as customers begin to demand performance guarantees. The most common reliability assessment methods utilize the following process: (1) they simulate a system’s response to a contingency, (2) they compute the reliability impact that this contingency has on each customer, (3) the reliability impact is weighted by the probability of the contingency occurring, and (4) steps 1–3 are repeated for all contingencies. Since this process results in the reliability that each customer can expect, new designs can be compared, existing systems can be analyzed, and reliability improvement options can be explored.

19.7 Distribution Reliability Indices Utilities typically keep track of customer reliability by using reliability indices. These are average customer reliability values for a specific area. This area can be the utility’s entire service area, a particular geographic region, a substation service area, a feeder service area, and so on. The most commonly used reliability indices give each customer equal weight. This means that a large industrial customer and a small residential

19-8

Power Systems

customer will each have an equal impact on computed indices. The most common of these customer reliability indices are as follows: System Average Interruption Frequency Index (SAIFI), System Average Interruption Duration Index (SAIDI), Customer Average Interruption Duration Index (CAIDI), and Average System Availability Index (ASAI) (IEEE, 2003). Notice that these indices are redundant. If SAIFI and SAIDI are known, both CAIDI and ASAI can be calculated. Formulae for these indices are as follows: SAIFI =

SAIDI =

∑ Customer interruption durations hours per year

CAIDI =

Total number of customer interruptions per year Total number of customers served

Total number of customers served

∑ Customer interruption durations Total number of customer interruptions

ASAI =

=

SAIDI hours per interruption SAIFI

Customer hours service availability 8760 − SAIDI = per unit 8760 Customer hours service demand

(19.3)

(19.4)

(19.5)

(19.6)

Some less commonly used reliability indices are not based on the total number of customers served. The Customer Average Interruption Frequency Index (CAIFI) and the Customer Total Average Interruption Duration Index (CTAIDI) are based upon the number of customers that have experienced one or more interruptions in the relevant year. The Average System Interruption Frequency Index (ASIFI) and the Average System Interruption Duration Index (ASIDI) are based upon the connected kVA of customers (these are sometimes referred to as load-based indices). Formulae for these indices are as follows: CAIFI = CTAIDI =

Total number of customer interruptions per year Customers experiencing one or more interruptions

∑ Customer interruption durations

Customers experiencing one or more interruptions

ASIDI =

(19.8)

Connected kVA interrupted per yeaar Total connected kVA served

(19.9)

Connected kVA hours interrupted ho ours per year Total connected kVA served

(19.10)

ASIFI =

hours per year

(19.7)

As momentary interruptions become more important, it becomes necessary to keep track of indices related to momentary interruptions. Since the duration of momentary interruptions is of little consequence, a single frequency-related index, the Momentary Average Interruption Frequency Index (MAIFI), is all that is needed. MAIFI, like SAIFI, weights each customer equally (there is currently no load-based index for momentary interruptions). The formula for MAIFI is as follows: MAIFI =

Total number of customer momentary interruptions per year Total number of customers served

(19.11)

19-9

Power System Reliability

The precise application of MAIFI varies. This variation is best illustrated by an example. Assume that a customer experiences three recloser operations followed by a recloser lockout, all within a period of 1 min. Some utilities would not count this event as a momentary interruption since the customer experiences a sustained interruption. Other utilities would count this event as three momentary interruptions and one sustained interruption. Similarly, if a customer experiences three recloser operations within a period of 1 min with power being restored after the last recloser, some utilities would count the event as three momentary interruptions and other utilities would count the event as a single momentary interruption. The IEEE defines a momentary interruption as lasting less than five minutes and a “momentary event” as a grouping of one or more momentary interruptions occurring within a five minute interval.

19.8 Storms and Major Events When electric utilities compute reliability indices, they often exclude interruptions caused by “storms” and “major events.” The definition of a major event varies from utility to utility, but a typical example is when more than 10% of customers experience an interruption during the event. The event starts when the notification of the first interruption is received and ends when all customers are restored service. The IEEE defines a major event based on a statistical approach called 2.5 Beta (IEEE, 2003). In nonstorm conditions, equipment failures are independent events—the failure of one device is completely independent of another device. In contrast, major events are characterized by common-mode failures. This means that a common cause is responsible for all equipment failures. The result is that many components tend to fail at the same time. This puts a strain on utility resources, which can only handle a certain number of concurrent failures (Brown et al., 1997). The most common causes of major events are wind storms, ice storms, and heat waves. Wind storms refer to linear winds that blow down trees and utility poles. The severity of wind storms is dependent upon sustained wind speed, gust speed, wind direction, and the length of the storm. Severity is also sensitive to vegetation management and the time elapsed since the last wind storm. Since a wind storm will tend to blow over all of the weak trees, a similar storm occurring a few months later may have little impact. A U.S. map showing wind speeds for the worst expected storm in 50 years is shown in Figure 19.4.

90(40) 100(45) 110(49)

85(38)

120(54) 90(40) 90(40) 130(58) 140(63)

140(63)

150(67)

140(63)

Special wind region 90(40) 100(45) 130(58) V mph (m/s) Location 110(49) 120(54) 105 (47) Hawaii 145 (65) Puerto Rico

FIGURE 19.4 Fifty-year wind storm (sustained wind speed in miles/h).

130(58) 140(63) 150(67)

19-10

Power Systems

Ice storms refer to ice buildup on conductors. This has four major effects: (1) it places a heavy physical load on the conductors and support structures, (2) it increases the cross-sectional area that is exposed to the wind, (3) ice can break off and cause a conductor to jump into the phase wires located above it, and (4) galloping. Galloping occurs when ice buildup assumes a teardrop shape and acts as an airfoil. During high winds, this can cause conductors to swing wildly and with great force. Ice can also cause problems by accumulating in trees, causing limbs to break off, and causing entire trunks to fall over into power lines. Heat waves are extended periods of exceedingly hot weather. This hot weather causes electricity demand to skyrocket due to air-conditioning loads. At the same time, conductors cannot carry as much electricity since they cannot transfer heat as effectively to their surroundings. This combination of heavy loading and conductor de-rating can cause overhead wires to become overloaded and sag to dangerous levels. Overloaded cables will cause insulation to lose life. In a worst-case scenario, the maximum power transfer capabilities of the system can be approached, resulting in a voltage collapse condition. Humidity exacerbates the impact of heat waves since it causes air conditioners to consume more energy.

19.9 Component Reliability Data For a reliability model to be accurate, component reliability data must be representative of the system being modeled. Utilities recognize this and are increasing their efforts to keep track of component failure rates, failure modes, repair times, switching times, and other important reliability parameters. Unfortunately, reliability statistics vary widely from utility to utility and from country to country. The range of equipment reliability data that can be found in published literature is shown in Table 19.2. Because component reliability is very system specific, it is beneficial to calibrate reliability models to historical reliability indices. In this process, component reliability parameters are adjusted until historical reliability indices match computed reliability indices (Brown and Ochoa, 1998). The amount that each parameter is adjusted should depend on the confidence of the original value and the sensitivity of the reliability indices to changes in this value. To illustrate, consider an overhead distribution system. A reliability model of this system is created using component reliability data from published literature. Unfortunately, the reliability indices that the model produces do not agree with the historical performance of the system TABLE 19.2 Equipment Reliability Data Component

Failure Rate (year−1)

MTTR (h)

Substation Equipment Power transformers Circuit breakers Disconnect switches Air-insulated buswork

0.015–0.07 0.003–0.02 0.004–0.16 0.002–0.04

15–480 6–80 1.5–12 2–13

Overhead Equipment Transmission linesa Distribution linesa Switches/fused cutouts Pole mounted transformer

0.003–0.140 0.030–0.180 0.004–0.014 0.001–0.004

4–280 4–110 1–4 3–8

Underground Equipment Cablea Padmount switches Padmount transformers Cable terminations/joints

0.005–0.04 0.001–0.01 0.002–0.003 0.0001–0.002

3–30 1–5 2–6 2–4

a

Failure rates for lines and cable are per mile.

Power System Reliability

19-11

over the past few years. To fix this, the failure rate and repair times of overhead lines (along with other component parameters) can be adjusted until predicted reliability matches historical reliability.

19.10 Utility Reliability Problems To gain a broader understanding of power system reliability, it is necessary to understand the root causes of system faults and system failures. A description of major failure modes is now provided.

19.10.1 Underground Cable A major reliability concern pertaining to underground cables is electrochemical treeing. Treeing occurs when moisture penetration in the presence of an electric field reduces the dielectric strength of cable insulation. When the dielectric strength is degraded sufficiently, transients caused by lightning or switching can result in dielectric breakdown. Electrochemical treeing usually affects extruded dielectric cable such as cross-linked polyethylene (XLPE) and ethylene-propylene rubber (EPR), and is largely attributed to insulation impurities and bad manufacturing. To reduce failures related to electrochemical treeing, a utility can install surge protection on riser poles (transitions from overhead to underground), can purchase tree-retardant cable, and can test cable reels before accepting them from the manufacturer. Existing cable can be tested and replaced if problems are found. One way to do this is to apply a DC voltage withstand test (approximately three times nominal RMS voltage). Since cables will either pass or not pass this test, information about the state of cable deterioration cannot be determined. Another popular method for cable testing is to inject a small signal into one end and check for reflections that will occur at partial discharge points. Other methods are measuring the power factor over a range of frequencies (dielectric spectroscopy), analyzing physical insulation samples in a lab for polymeric breakdown (degree of polymerization), and using cable indentors to test the hardness of the insulation. Not all underground cable system failures are due to cable insulation. A substantial percentage occurs at splices, terminations, and joints. Major causes are due to water ingress and poor workmanship. Heat shrink covers can be used to waterproof these junctions and improve reliability. The last major reliability concern for underground cable is dig-ins. This is when excavation equipment cuts through one or more cables. To prevent dig-ins, utilities should encourage the public to have cable routes identified before initiating site excavation. In extreme cases where high reliability is required, utilities can place cable in concrete-encased duct banks.

19.10.2 Transformer Failures Transformers are critical links in power systems, and can take a long time to replace if they fail. Throughfaults cause extreme physical stress on transformer windings, and are the major cause of transformer failures. Overloads rarely result in transformer failures, but do cause thermal aging of winding insulation. When a transformer becomes hot, the insulation on the windings slowly breaks down and becomes brittle over time. The rate of thermal breakdown approximately doubles for every 10°C. Ten degree Celcius is referred to as the “Montsinger Factor” and is a rule of thumb describing the Arrhenius theory of electrolytic dissociation. Because of this exponential relationship, transformer overloads can result in rapid transformer aging. When thermal aging has caused insulation to become sufficiently brittle, the next fault current that passes through the transformer will mechanically shake the windings, a crack will form in the insulation, and an internal transformer fault will result. Extreme hot-spot temperatures in liquid-filled transformers can also result in failure. This is because the hot spot can cause free bubbles that reduce the dielectric strength of the liquid. Even if free bubbles are not formed, high temperatures will increase internal tank pressure and may result in overflow or tank rupture. Many transformers are fitted with load tap changers (LTCs) for voltage regulation. These mechanically moving devices have historically been prone to failure and can substantially reduce the reliability

19-12

Power Systems 10 20

20 30 40 20 70 50

50 30

60

40 50

60 10

60 70

70

20 30

40

70 60 80 50

90

100 80 70 60

FIGURE 19.5 Number of thunderstorm days per year.

of a transformer (Willis, 1997). Manufacturers have addressed this problem and new LTC models using vacuum technology have succeeded in reducing failure rates.

19.10.3 Lightning A lightning strike occurs when the voltage generated between a cloud and the ground exceeds the dielectric strength of the air. This results in a massive current stroke that usually exceeds 30,000 A. To make matters worse, most strokes consist of multiple discharges within a fraction of a second. Lightning is the major reliability concern for utilities located in high keraunic areas (Burke, 1994). An isokeraunic map for the United States is shown in Figure 19.5. Lightning can affect power systems through direct strikes (the stroke contacts the power system) or through indirect strikes (the stroke contacts something in close proximity and induces a traveling voltage wave on the power system). Lightning can be protected against by having a high system BIL (basic impulse level) by using shield wires, by using surge arrestors to clamp voltages across equipment, and by having a low impedance ground. Direct strikes are virtually impossible to protect against on a distribution system.

19.10.4 Tree Contact Trees continuously grow, can fall over onto conductors, can drop branches onto conductors, can push conductors together, and can serve as gateway for animals. This is why many utilities spend more on tree trimming than on any other preventative maintenance activity. When a tree branch bridges two conductors, a fault does not occur immediately. This is because a moist tree branch has a substantial resistance. A small current begins to flow and starts to dry out the wood fibers. After several minutes, the cellulose will carbonize, resistance will be greatly reduced, and a short circuit will occur. Branches brushing against a single phase conductor typically do not result in system faults. Faults due to tree contact can be reduced by using tree wire. This is overhead wire with an insulated jacket similar to cable. Tree wire can be effective, but faults tend to result in conductor burndown since they will not motor (move themselves along the conductor) like faults on bare conductor.

Power System Reliability

19-13

19.10.5 Birds Birds are the most common cause of animal faults on both transmission systems and air-insulated substations. Different types of birds cause different types of problems, but they can generally be classified as nesting birds, roosting birds, raptors, and woodpeckers. Nesting birds commonly build their homes on transmission towers and in substations. Nesting materials can cause faults, and bird excrement can contaminate insulators. Nesting birds also attract predators such as raccoons, snakes, and cats. These predators can be a worse reliability problem than the birds themselves. Roosting birds use electrical equipment to rest on or to search for prey. They can be electrocuted by bridging conductors with their wings, and their excrement can contaminate insulators. To prevent birds from roosting, anti-roosting devices can be placed on attractive sites. For locations that cater to thousands of roosting birds, more extreme deterrent methods such as pyrotechnics can be used. Raptors are birds of prey such as eagles, hawks, ospreys, owls, and vultures. Reliability problems are similar to other roosting and nesting birds, but special consideration may be required since most raptors are protected by the federal government. Woodpeckers peck holes in wood with their beaks as they search for insects. This does not harm trees (the bark regenerates), but can cause devastating damage to utility poles. This can be prevented by using steel poles, by using repellent, or by tricking a woodpecker into believing that there is already a resident woodpecker (woodpeckers are quite territorial).

19.10.6 Squirrels Squirrels are a reliability concern for all overhead distribution systems near wooded areas. Squirrels will not typically climb utility poles, but will leap onto them from nearby trees. They cause faults by bridging grounded equipment with phase conductors. Squirrel problems can be mitigated by cutting down nearby access trees or by installing animal guards on insulators.

19.10.7 Snakes Snakes are major reliability concerns in both substations and underground systems. They can squeeze through very small openings, can climb almost anything, and have the length to easily span phase conductors. Snakes are usually searching for food (birds in substations and mice in underground systems), and removing the food supply can often remove the snake problem. Special “snake fences” are also available.

19.10.8 Insects It is becoming more common for fire ants to build nests in pad-mounted equipment. Their nesting materials can cause short circuits, the ants can eat away at conductor insulation, and they make equipment maintenance a challenge.

19.10.9 Bears, Bison, and Cattle These large animals do not typically cause short circuits, but degrade the structural integrity of poles by rubbing on guy wires. Bears can also destroy wooden poles by using them as scratching posts, and black bears can climb wooden utility poles. These problems can be addressed by placing fences around poles and guy wire anchors.

19-14

Power Systems

19.10.10 Mice, Rats, and Gophers These rodents cause faults by gnawing through the insulation of underground cable. They are the most common cause of animal-related outages on underground equipment. To make matters worse, they will attract snakes (also a reliability problem). Equipment cabinets should be tightly sealed to prevent these small animals from entering. Ultrasonic devices can also be used to keep rodents away (ultrasonic devices will not keep snakes away).

19.10.11 Vandalism Vandalism can take many different forms, from people shooting insulators with rifles to professional thieves stealing conductor wire for scrap metal. Addressing these reliability problems will vary greatly from situation to situation.

19.11 Reliability Economics When a power interruption occurs, both the utility and the interrupted customers are inconvenienced. The utility must spend money to fix the problem, will lose energy sales during the interruption, and may be sued by disgruntled customers. From the customer perspective, batch processes may be ruined, electronic devices may crash, production may be lost, retail sales may be lost, and inventory (such as refrigerated food) may be ruined. When a customer experiences an interruption, there is an amount of money that it would be willing to pay to have avoided the interruption. This amount is referred to as the customer’s incurred cost of poor reliability, and consists of a base cost plus a time-dependent cost. The base cost is the same for all interruptions, relates to electronic equipment shutdown and interrupted processes, and is equivalent to the cost of a momentary interruption. The time-dependent cost relates to lost production and extended inconvenience, and reflects that customers would prefer interruptions to be shorter rather than longer. The customer cost of an interruption varies widely from customer to customer and from country to country. Other important factors include the time of year, the day of the week, the time of day, and whether advanced warning is provided. Specific results are well documented by a host of customer surveys (Billinton et al., 1983; IEEE Std. 493-1990; Tollefson et al., 1991, 1994). For planning purposes, it is useful to aggregate these results into a few basic customer classes: commercial, industrial, and residential. Since larger customers will have a higher cost of reliability, results are normalized to the peak kW load of each customer. Reliability cost curves for typical U.S. customers are shown in Figure 19.6.

Customer cost ($/kW)

50 40

C

30

10

l

stria

Indu

20

0

ial

erc

m om

Residential 0

1

2 3 Duration of interruption (h)

FIGURE 19.6 Typical U.S. customer interruption costs (1999 dollars).

4

Power System Reliability

19-15

Average customer cost curves tend to be linear and can be modeled as an initial cost plus a first order, time-dependent cost. Specific customer cost curves may be extremely nonlinear. For example, a meat packing warehouse depending upon refrigeration may be unaffected by interruptions lasting many hours. At a certain point, the meat will begin to spoil and severe economic losses will quickly occur. After the meat spoils, additional interruption time will harm this particular customer much more.

19.12 Annual Variations in Reliability Power system reliability varies from year to year. In a lucky year, a system may have a SAIDI of 30 min. The next year, this exact same system may experience a SAIDI of 8 h. This type of variation is inevitable and must be considered when comparing reliability indices. It is also important to note that the variance of reliability indices will tend to be less for areas serving more customers. Individual customer reliability will tend to be the most volatile, followed by feeder reliability, substation reliability, regional reliability, and so forth. The importance of annual reliability variance will grow as utilities become subject to performancebased rates and as customer reliability guarantees become more common. These types of contracts expose utilities to new risks that must be understood and managed. Since performance-based contracts penalize and reward utilities based on reliability, annual variations must be understood for fair contracts to be negotiated and managed. Contractual issues concerning service reliability are becoming important as the electric industry becomes more competitive. Customers can choose between suppliers, and realize that there is a tradeoff between reliability and rates. Some customers will demand poor reliability at low rates, and other customers will demand high reliability at premium rates. To address the wide variation in customer needs, utilities can no longer be suppliers of energy alone, but must become suppliers of both energy and reliability. Power system reliability is now a bona fide commodity with explicit value for utilities to supply and explicit value that customers demand.

References Allan, R.N., Billinton, R., Breipohl, A.M., and Grigg, C.H., Bibliography on the application of probability methods in power system reliability evaluation, IEEE Trans. Power Syst., 9, 1, 41–49, Feb. 1994. Billinton, R. and Allan, R.N., Reliability Assessment of Large Electric Power Systems, Kluwer Academic Publishers, Dordrecht, the Netherlands, 1988. Billinton, R. and Allan, R., Reliability Evaluation of Engineering Systems: Concepts and Techniques, 2nd edn., Plenum Press, New York, 1992. Billinton, R. and Jonnavitihula, S., A test system for teaching overall power system reliability assessment, IEEE Trans. Power Syst., 11, 4, 1670–1676, Nov. 1996. Billinton, R., Wacker, G., and Wojczynski, E., Comprehensive bibliography on electrical service interruption costs, IEEE Trans. Power Appar. Syst., PAS-102, 6, 1831–1837, June 1983. Brown, R.E., Electric Power Distribution Reliability, Second Edition, CRC Press, 2009. Brown, R.E., Gupta, S., Christie, R.D., Venkata, S.S., and Fletcher, R.D., Distribution system reliability analysis using hierarchical Markov modeling, IEEE Trans. Power Delivery, 11, 4, 1929–1934, Oct. 1996. Brown, R.E., Gupta, S., Christie, R.D., Venkata, S.S., and Fletcher, R.D., Distribution system reliability: Momentary interruptions and storms, IEEE Trans. Power Delivery, 12, 4, 1569–1575, Oct. 1997. Brown, R.E. and Ochoa, J.R., Distribution system reliability: Default data and model validation, IEEE Trans. Power Syst., 13, 2, 704–709, May 1998. Burke, J.J., Power Distribution Engineering, Marcel Dekker, Inc., New York, 1994. Electrical Transmission and Distribution Reference Book, ABB Power T&D Company, Inc., Raleigh, NC, 1997.

19-16

Power Systems

Endrenyi, J., Reliability in Electric Power Systems, John Wiley & Sons, Ltd., New York, 1978. IEEE Std. 493-1990, IEEE Recommended Practice for the Design of Reliable Industrial and Commercial Power Systems, 1990. IEEE Standard 1366–2003, Guide for Electric Power Distribution Reliability Indices, 2003. Philipson, L. and Willis, H.L., Understanding Electric Utilities and De-regulation, Marcel Dekker, Inc., New York, 1999. Ramakumar, R., Engineering Reliability: Fundamentals and Applications, Prentice-Hall, Inc., Englewood Cliffs, NJ, 1993. Schilling, M.T., Billinton, R., Leite da Silva, A.M., and El-Kady, M.A., Bibliography on composite system reliability (1964–1988), IEEE Trans. Power Syst., 4, 3, 1122–1132, Aug. 1989. Tollefson, G., Billinton, R., and Wacker, G., Comprehensive bibliography on reliability worth and electric service consumer interruption costs 1980–1990, IEEE Trans. Power Syst., 6, 4, 1508–1514, Nov. 1991. Tollefson, G., Billinton, R., Wacker, G., Chan, E., and Aweya, J., A Canadian customer survey to assess power system reliability worth, IEEE Trans. Power Syst., 9, 1, 443–450, Feb. 1994. Willis, H.L., Power Distribution Planning Reference Book, Marcel Dekker, Inc., New York, 1997.

20 Probabilistic Methods for Planning and Operational Analysis 20.1 20.2 20.3 20.4 20.5 20.6 20.7

Gerald T. Heydt Arizona State University

Peter W. Sauer University of Illinois at Urbana-Champaign

Uncertainty in Power System Engineering..................................20-1 Deterministic Power Flow Studies................................................20-2 Monte Carlo Power Flow Studies..................................................20-3 Analytical Probabilistic Power Flow Studies............................. 20-4 Applications for Available Transfer Capability (ATC)............. 20-6 An Example of Stochastic ATC.................................................... 20-8 An Example of Expected Financial Income from Transmission Tariffs............................................................. 20-8 20.8 Wind Energy Resources and Stochastic Power Flow Studies...... 20-10 20.9 Solar Photovoltaic Energy Resources and Stochastic Power Flow Studies........................................................................ 20-11 20.10 Conclusions.................................................................................... 20-11 References................................................................................................... 20-11

20.1 Uncertainty in Power System Engineering Probabilistic methods are mathematical techniques to formally consider the impact of uncertainty in models, parameters, or data. Typical uncertainties include the future value of loading conditions, fuel prices, weather, and the status of equipment. Methods to consider all possible values of uncertain data or parameters include such techniques as interval analysis, minimum/maximum analysis, and fuzzy mathematics. In many cases, these techniques will produce conservative results because they do not necessarily incorporate the “likelihood” of each value of a parameter in an expected range, or they might be intentionally designed to compute “worse case” scenarios. These techniques are not discussed further here. Instead, this chapter presents two techniques that have been successfully applied to power system planning and operational analysis as noted by the “Application of Probability Methods” subcommittee of the IEEE Power Engineering Society in Rau et al. (1994). They are • Monte Carlo simulation: analysis in which the system to be studied is subjected to pseudorandom operating conditions, and the results of many analyses are recorded and subsequently statistically studied. The advantage is that no specialized forms or simplifications of the system model are needed. • Analytical probability methods: analysis in which the system to be studied is represented by functions of several random variables of known distribution. Rau et al. (1994) also noted that probabilistic load flow methods would be well suited to evaluate loadability limits and transfer capabilities under uncertainties created by industry restructuring. They also noted the 20-1

20-2

Power Systems

needs to include uncertainty analysis in cost/worth studies and security assessment. This chapter focuses on those two analysis tools and illustrates the differences between the Monte Carlo simulation method and the analytical probability method. There has been renewed interest in the subject of uncertainty in generation with the incorporation of renewable resources. Mainly these resources are wind energy, solar energy, and concentrated solar energy. In each of these cases, uncertainty in the supply power occurs because of the variability of the wind and the stochastic nature of wind forecasts, as well as the variability of insolation (solar intensity) due to weather and man-made degradation to atmospheric transmissivity. One approach in these areas is to “bracket” the variability of the generation (e.g., high wind case, high solar case versus low wind case, low solar case). Another is to model the probability density function of the wind or solar generation, and then use various algorithms that process the probability density functions of the sources to obtain the probability density functions of the line loads and bus voltages (i.e., a “stochastic load flow study”). The stochastic analysis described makes available such information as the probability of insufficient generation, the probability of high line loads (e.g., overload), and the probability of high bus voltage (this may occur during light load conditions with high penetration of solar generation at unity power factor common to residential solar photovoltaic generation systems).

20.2 Deterministic Power Flow Studies Conventional, deterministic power flow studies (load flow studies) are one of the most widely used analysis tools for both planning and operations (Heydt, 1996). They can imprecisely be described through three features: • Power flow analysis is the computation of steady-state conditions for a given set of loads P (active power) and Q (reactive power) and a given system configuration (interconnecting lines/transformers). • Active-power generation and bus voltage magnitudes are specified at the generator buses. • Power flow analysis gives a solution of system states, mainly bus voltage phase angles (δ) and magnitudes (V).

Deterministic, single-valued numbers

Additional complexities such as phase-shifting and tap-changing-under-load (TCUL) transformers and other devices are often added to this basic description. And of course many other “solution” quantities such as generator reactive power output and line power flows are often added. The main method used for power flow analysis is the Newton–Raphson (NR) method, a method to find the zeros of f(X) = 0 where f is a vector-valued function of P and Q summations at each bus (Kirchhoff’s current law in terms of power), and X is the system state vector (primarily δ and V). There is normally a natural weak coupling between real power and voltage magnitude (P and V) as well as reactive power and voltage angle (Q and δ). This weak coupling can be exploited in the NR method to speed up solutions. Figures 20.1 and 20.2 are pictorials of this decoupling.

Load data (P and Q)

Generation data (P and V)

System configuration (Interconnecting lines)

FIGURE 20.1 Conventional NR power flow study.

Full NewtonRaphson solves the P and Q summation equations and updates step for δ and V

Residual check and output δ and V

20-3

Deterministic, single-valued numbers

Probabilistic Methods for Planning and Operational Analysis

Load and generation data (P) System configuration (interconnecting lines)

Load data (Q) and generation data (V)

Decoupled NewtonRaphson solves the P summation equations and updates step for δ Decoupled NewtonRaphson solves the Q summation equations and updates step for V

Residual check and output δ and V

FIGURE 20.2 Decoupled NR power flow study.

This decoupling can be utilized to increase the speed of computing exact solutions, and also can be used to create superfast approximate linear solutions. In both cases, the inputs to the power flow solution algorithms shown are deterministic. When there is uncertainty in the input data, this can be considered using the two methods introduced earlier—Monte Carlo and analytical probability. These two methods are discussed next in the context of power flow analysis.

20.3 Monte Carlo Power Flow Studies In Monte Carlo power flow studies, multiple power flow studies are run using different sets of input data. The various sets of input data are typically obtained from random number generators to correspond to some desired statistical distribution. For example, the true future total load of a power system (and its allocation to system buses) might be unknown, but an expected value might be known together with a variance with specified statistical distribution (i.e., uniform or normal). This problem could be a problem with a scalar random variable (random total load at one time instant with known deterministic allocation to system buses), or it could be a problem with a vector of random variables (random total load at one time instant plus random allocation to system buses). In the scalar case, this would require numerous power flow studies using numerous random samples for the total load. In the vector case, this would require numerous power flow studies using numerous random samples for the total load (perhaps one statistical distribution) plus numerous random samples for the factors that allocate the total load to the system buses (perhaps another statistical distribution). This could perhaps be defined more easily by simply considering each system bus load to be a random variable. This would require a decision of which “multivariate” statistical distribution to use (i.e., all bus load values are independent and normally distributed, or all bus loads are fully correlated and normally distributed, or something in between). In most cases, the Monte Carlo power flow studies are done for “variations” from the expected value of the random variables. As such, the expected value of the variation in bus load from the expected value would be zero. The first Monte Carlo power flow study is usually the case where all variations are equal to their expected value (zero). This is also often called the “base case.” Figure 20.3 shows the Monte Carlo simulation method for considering uncertainty in the input data. The statistical description of the distribution is typically a histogram. This is a plot of the value of the output variable on the horizontal axis and the number of times that value occurs on the vertical axis. The horizontal values are grouped into “ranges” of possible values (i.e., all values from 0.1 to 0.2, 0.2 to 0.3, etc.). These histograms are similar in concept to the probability density function of analytical methods. There are also numerical descriptions (moments) of the statistical distribution that can be computed—i.e., mean, variance, skewness, kurtosis, etc. One problem with the Monte Carlo method is in determining the number of samples that are sufficient to properly represent the variation in the

20-4

Pseudorandom sample of numbers

Power Systems

Load data (P and Q)

Generation data (P and V)

System configuration (interconnecting lines)

Full Newton– Raphson solves the P and Q summation equations and updates step for δ and V

Residual check and output δ and V

Add to the statistical description of the distribution of output quantities. Stop when enough

FIGURE 20.3 Monte Carlo power flow study.

uncertain parameters. For example, are 100 samples sufficient to represent all possible variations in the total system load? How about 1000? One way to look at this is—“The deterministic solution is just one sample, so anything more than one is better than the deterministic solution.” But, what is important in the choice of samples is the specification of what is most likely to represent the most likely possible values of the uncertain data. This normally requires a large number of samples. Another problem with Monte Carlo methods is the large computational burden required to study a large number of samples. Multiple power flow studies can also be used to compute the variation of system bus voltages (and all other output quantities) throughout the day, week, or year as the load varies. For example, if there was an interest in computing the line flows for every hour of the day for 1 year (assuming all load and generation data are known for each hour), this would require 8760 power flow studies (1/h). It is important to realize that although this is not really a Monte Carlo simulation as the variation of the load data is not from a random phenomenon (i.e., daily, weekly, and seasonal variations in load are somewhat predictable), it is computationally the same as running 8760 “samples” and observing the distribution of the output quantities. The “statistical distribution” would give an indication of the percentage of time throughout the year during which a line flow was in a certain range. If there was an interest in considering the uncertainty in the total load variation throughout the year, some mechanism for specifying the “variations” from the base case would need to be given for each hour of everyday. Presumably, these variations would take the form of random samples created from some known statistical distribution. A Monte Carlo simulation of this could easily result in the need for over one million power flow studies. And, if the uncertainty of the data was extended to random samples of load at each bus for each hour, the result could be the need for over one billion power flow studies.

20.4 Analytical Probabilistic Power Flow Studies Because it is very difficult to perform analytical probability analysis for large nonlinear systems, the most probabilistic power flow studies exploit two properties of power systems that lead to a linear, smaller model. First, the decoupling that was discussed above allows the real power and voltage angle to be completely separated from the reactive power and voltage magnitude in computation. This not only reduces the size of the problem, but also reduces the accuracy and loses voltage information. Second, the fact that “small” changes in bus real power injections result in “small” changes in bus voltage angles leads to a linear approximation as follows:

∆δ bus = T1∆Pbus

(20.1)

So, for an initial power flow solution giving the bus voltage angle vector δ0bus for an initial bus real power 1 0 0 = Pbus + ∆Pbus is given injection vector, Pbus , an approximate power flow solution for a problem with Pbus 1 0 by δbus = δbus + T1∆Pbus .

Probabilistic Methods for Planning and Operational Analysis

20-5

It is important to point out that the vector of bus real power injection changes does not include the injection change at the swing (or slack) bus of the power flow study. And, the vector of bus voltage angles does not include an entry for the swing (or slack) bus of the power flow study. This is because the swing (or slack) bus is chosen as the fixed-angle reference for the power flow study and also provides the balance in real power for any specified condition. In this manner, the vector of bus real power injection changes could include a single entry and the result of this injection change would automatically be offset by a nearly equal and opposite injection change at the swing bus. This linear analysis can be extended to approximate the resulting change in real power line flows as

∆Pline = T2 ∆δ bus

(20.2)

This combination of linear approximations gives the traditional “generation shift distribution factor,” which is also the “power transfer distribution factor” for power transfers from a bus to or from the swing bus (Heydt, 1996):

∆Pline = T2T1∆Pbus = T ∆Pbus

(20.3)

This relationship is the heart of the linear power flow, which enables analytical probabilistic methods. One reason a linear computation is important is that the linear sum of jointly normal random variables is also jointly normal (Papoulis, 1965). This makes the computation of the statistical distribution of the output variables very easy to compute from the specified statistical distribution of the input variables. For example, if there is some uncertainty in the value of the loads, an initial power flow study could be done using the expected value of the loads. This would be a deterministic power flow study. To account for the uncertainty in load, an analytical probabilistic power flow could be done as follows: Step 1. Specify the statistical distribution of the variation of load (ΔP bus)—for a scalar (total load) this might be a zero mean, normal distribution with some specified variance. For a vector of zero-mean jointly normal load variations, the statistical distribution would be specified by also giving the covariance between each load variation. A covariance of zero would mean that the two uncertain jointly normal variations are statistically “independent.” A covariance of ±1.0 would mean that the two jointly normal variations are fully correlated (linearly related). The covariance matrix C load contains the complete statistical description of all the load variations through the variances (diagonal) and covariances (off diagonal). Step 2. Compute the statistical distribution of the variation of line real power flows. When the variations in loads are assumed to be zero-mean jointly normal, the complete description of the statistical distribution of the variation of line real power flows is found easily as (Papoulis, 1965)

Cline = T3CloadT3t

(20.4)

where “t” denotes matrix transposition. The mean value of the variation in line real power flows is taken as zero because the mean value of the bus real power injection variations was assumed zero and the two are assumed to be linearly related. Now, the fact that the variations in line real power flows are not really linearly related to the bus real power injection variations means that these statistical distributions will not be exact. Figure 20.4 shows how the analytic probabilistic power flow study is done. Stochastic power flow is another term that has been used for analytical probabilistic power flow. Since stochastic processes are statistical processes involving a number of random variables depending on a variable parameter (usually time), this terminology has been adopted as equivalent. One of the first publications on this method was of Borkowska (1974). The subject continues to attract interest (Vorsic et al., 1991).

20-6

Power Systems

Load and generation statistical distribution data (P) System configuration (interconnecting lines)

Linear power flow solution for the statistical distribution of δ and other output variables

FIGURE 20.4 Analytic probabilistic power flow study.

20.5 Applications for Available Transfer Capability (ATC) In order to realize open access to electric power transmission networks and promote generation competition and customer choice, the Federal Energy Regulatory Commission requires that ATC be made available on a publicly accessible open access same-time information system (OASIS). The ATC is defined as a measure of the transfer capability, or available “room” in the physical transmission network for transfers of power for further commercial activity, over and above already committed uses (NERC, 1996). The ATC is defined by NERC as the total transfer capability (TTC) minus the transmission reliability margin (TRM), capacity benefit margin (CBM), and existing power flows:

ATC = TTC − TRM − CBM − Existing power flows

(20.5)

where the TTC is the total amount of power that can be sent from bus A to bus B within a power network in a reliable manner, the TRM is the amount of transmission transfer capability necessary to ensure network security under uncertainties, and the CBM is the amount of TTC reserved by load serving entities to ensure access to generation from interconnected systems to meet generation reliability requirements. Neglecting TRM and CBM reduces the computation of ATC to ATC = TTC − Existing power flows.

(20.6)

The ATC between two buses (or groups of buses) determines the maximum additional power that can be transmitted in an interchange schedule between the specified buses. ATC is clearly determined by load flow study results and transmission limits. As an illustration, consider Figure 20.5 in which 1000 MW is transmitted from bus A to bus B. If the line rating is 1500 MW and if no parallel paths exist from A to B, the ATC from A to B is 500 MW. In a real power system, the network and the computation of the ATC are much more complicated as parallel flows make the relationship between transfers and flows less obvious. In addition, the ATC is uncertain due to the uncertainty of power system equipment availability and power system loads. An evaluation of the stochastic behavior of the ATC is important to reduce the likelihood of congestion. The ATC is determined by power flow studies subject to transmission system limits. These phenomena are Line rating 1500 MW (ATC = 500 MW) B

A P = 1000 MW

FIGURE 20.5 An example of ATC.

Probabilistic Methods for Planning and Operational Analysis

20-7

nonlinear in behavior. As discussed above, linearization can be used to estimate line flows in a power flow study model. This section presents a method of finding the stochastic ATC using a linear transformation of line power flows into ATC. Increasing the transfer power increases the loading in the network. At some point, operational or physical limits to various elements are reached that prevent further increase. The largest value of transfer that causes no limit violations is used to compute the TTC and ATC. Limits are affected by the power injections at both buses A and B. This effect can be found analytically by finding the distribution factors of the lines and other components. In this context, a “distribution factor” refers to the power transfer distribution factor for line i − j and bus k, where an equal and opposite injection is automatically made at the swing bus. The complexity of the ATC calculation is drastically reduced by linearizing the power flow study problem and considering only thermal transmission limits. The linearization is most accurate when only small deviations from the point of linearization are encountered. The use of thermal limits is justified for short transmission circuits. The transmission system linearization in this case is done using power transfer distribution factors, which are discussed above. Starting with a base case power flow solution, the addition of a power transfer from bus A to bus B impacts the flow in the line between bus i and bus j as

Pij1 = Pij0 + (Tij , A − Tij ,B )PAB

(20.7)

where PAB is the power transfer from bus A to bus B (assumed to be positive). When (Tij,A − Tij,B) is positive, the transfer that results in rated power flow on line i − j is

PAB,ij =

(P

Rated ij

− Pij0

)

(Tij , A − Tij ,B )

(20.8)

When (Tij,A − Tij,B) is negative, the transfer that results in rated power flow on line i − j is

PAB,ij =

( −P

Rated ij

− Pij0

(Tij , A − Tij ,B )

)

(20.9)

Each line will have a value of PAB,ij that represents the maximum value of power transfer from bus A to bus B without overloading line ij. The minimum of all these values would be the ATC for the AB transfer. Once this minimum is determined, the question is, “What if there is uncertainty in the initial bus and line loadings?” As discussed above, there are at least two ways to incorporate uncertainty in this computation. The first way is the Monte Carlo method that specifies a statistical distribution of the uncertain parameter (perhaps total load) and utilizes pseudorandom number samples to perform brute force repeated ATC solutions. This could involve full nonlinear power flow studies, or faster linear power flow studies. The collection of ATC solutions that come from a large sample of total load values gives the statistical distribution of the ATC, perhaps as a histogram. If the critical line that determines the ATC of the AB transfer does not change with the uncertainty of the initial bus and line loadings, then the possible variation in ATC can be easily computed from the statistical distribution of the initial line flow. For example, if the uncertainty of initial line flows is normally distributed with zero mean and some variance, then the ATC from A to B is also normally distributed with zero mean, as it is a linear function of the initial flow in the critical line. The variance is also directly available because of the linear transformation.

20-8

Power Systems TABLE 20.1 Comparison of Monte Carlo and Analytical Methods for ATC “From” Bus 1 1 1 2

“To” Bus

Analytical Mean

Analytical Variance

9 10 11 9

8.5 4.177 2.7144 7.8480

4.5 2.3005 0.7038 4.5208

Monte Carlo Mean 7.147 4.7766 2.764 7.318

Monte Carlo Variance 5.008 2.3356 1.1549 5.6584

20.6 An Example of Stochastic ATC The following example illustrates the basic principles of stochastic ATC. The IEEE 14-bus system was studied with the base case bus and line variables as given in Christie (1999) and Pai (1979). For the uncertainty analysis by Monte Carlo methods, the bus loads were generated on the computer using 100 samples from a random number generator. The bus real power loads were assumed to be statistically independent with a normal distribution that had a mean equal to the base case and a standard deviation equal to 10% of the base case. Bus load power factor was maintained at the base case value. A comparison of the Monte Carlo results with the analytical method is given in Table 20.1. The Monte Carlo results required 100 full nonlinear power flow studies of ATC, whereas the analytical results required a simple direct multiplication. Additional results for this example are available in Stahlhut et al. (2005).

20.7 A n Example of Expected Financial Income from Transmission Tariffs A second illustration of the two methods for considering uncertainty deals with the estimation of transmission tariff revenue because of unknown future loading. This illustration was reported also in Westendorf (2005) and Stahlhut et al. (2005). The work is presented in three parts. The first includes a study of transmission tariff revenue based on an estimated hourly forecast for 1 year. This “base case” solution provides the expected revenue for the year for that forecasted loading level. The second part includes a Monte Carlo simulation of possible variations in revenue for a forecast error of known statistical distribution. This is a repeat of the base case solution for 100 different possible cases (per hour) sampled from a normal distribution function—using total load as the random variable. This error in total load was then “distributed” to the individual buses using the base case percentage allocation. The third part included an analytical solution to estimate the expected statistical distribution of the transmission tariff revenue using linear power flow techniques. This analysis used the IEEE 14-bus test system (Pai, 1979; Christie, 1999) and computed revenue from transmission tariffs using an assumed tariff of $0.04 per MW-mile for each transmission line. An annual income was also computed by summing the hourly incomes for the year. Because the transmission tariff is a function of the length of each transmission line, an estimate of the transmission line lengths was necessary. The total impedance of a transmission line is dependent on the length of the line. The line length is directly proportional to the transmission line impedance. The method for calculating the lengths of the transmission lines in the IEEE 14-bus test system based on the total line impedance of each line was developed as follows. A power base for the system was assumed to be Sbase = 100 MVA. Using the nominal voltages, Vbase, of the transmission lines and the power base of the system, each transmission line reactance value, the length of the line was calculated by assuming a conversion factor of 0.7 Ω/mile. Secondly, lines containing transformers (i.e., lines 8–10) were considered to be zero-length lines. A load forecast taking into account hourly variations was developed for this investigation by using publicly accessible approximate historical load data for each hour of everyday for the year 2004 from

20-9

Probabilistic Methods for Planning and Operational Analysis TABLE 20.2 Comparison of Monte Carlo and Analytical Methods for Expected Revenue for 1 Day Time

Mean Total Revenue ($)

Base Case Total Revenue ($)

Simulated Standard Deviation σrevenue ($)

Analytical Standard Deviation σrevenue ($)

100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300 2400

390.66 382.23 378.27 383.13 394.60 427.99 478.56 510.09 505.21 501.30 497.15 490.09 494.68 488.98 484.86 484.97 502.80 537.99 543.86 536.59 527.99 500.62 464.40 427.75

390.96 382.95 381.11 384.06 393.56 426.22 479.04 510.85 506.51 501.86 496.49 491.41 490.73 488.53 483.91 484.78 504.46 538.33 544.99 538.02 527.33 502.73 466.21 430.47

12.49 11.28 12.82 12.29 13.23 14.47 14.23 16.94 14.03 13.83 13.93 12.79 13.83 15.66 15.24 14.77 14.16 17.39 16.76 19.71 14.94 15.19 13.47 14.41

9.96 9.76 9.72 9.79 10.03 10.83 12.13 12.91 12.81 12.69 12.56 12.43 12.42 12.36 12.25 12.27 12.76 13.59 13.75 13.58 13.32 12.71 11.81 10.93

the PJM Web site (PJM, 2005). The integrated hourly load data for the PJM-E area were scaled to match the system load of 259 MW of the IEEE 14-bus test system. Using the hourly load data, power flows were conducted to determine the megawatt flows on the transmission lines allowing the calculation of the each line revenue for the hour. The total revenue generated for each hour was calculated by summing all of the line revenues calculated from the power flows. While standard nonlinear power flow methods were used to compute the Monte Carlo solutions, standard linear power flow methods were used to calculate the analytical solutions. The Monte Carlo simulations and the analytical methods were compared. Both methods provided results for the distribution of the system revenue generated for each hour. Although an entire year was analyzed, Table 20.2 lists the results from both methods for just 1 day. Column 1 is the time of day studied. Column 2 is the expected revenue using the Monte Carlo simulation, which created deviations from the base case using a zero-mean random number generator (100 samples per hour). Column 3 shows the revenue from the base case. The fact that these two columns are quite close is due to the fact that the deviations were sampled from a zero-mean number generator and the system behavior is nearly linear for deviations of this size. The standard deviation of the sampled distribution was 3%. This was designed to produce a sample load forecast with expected variations between plus and minus 10% of the base case. Column 4 shows the Monte Carlo sample standard deviation, and Column 5 shows the analytical standard deviation found using the assumed linear nature of the calculation and the assumed normal distribution of the forecast error.

20-10

Power Systems

The error in standard deviations can be attributed to several things. The linear power flow method is an approximation that is used in the analytical method but not in the Monte Carlo simulation method. The Monte Carlo method only used 100 random samples to create the sample standard deviation. This is not a large number of samples and contributes to the discrepancy in the two methods.

20.8 Wind Energy Resources and Stochastic Power Flow Studies Wind is caused by regions on Earth that have different air pressure. Perez (2007) and Pavia (1986) are brief samples of the extensive meteorological literature in this area. The wind speed is often characterized as a random process, and one particular model used frequently is the Weibull distribution (see Figure 20.6) because this distribution is one sided (i.e., the random variable is always non-negative), and actual wind data appear to behave in this fashion. The Weibull probability density f(x) is kx k −1e −( x /λ )k f x (x ) = λk 0

for x ≥ 0 otherwise

where λ is the scale of the density k is termed the shape parameter

Probability density

The actual power contained in wind is proportional to the cube of the wind speed under some conditions, and therefore the available wind power, if viewed as a random process, does not possess a Weibull probability density. The power output of a wind turbine, however, is not simply characterized because the cubic behavior of the wind power with wind speed only holds below a certain threshold of wind speed (typically in the