T&D System Design and Construction for Enhanced Reliability and Power Quality

Technical Report

Effective December 6, 2006, this report has been made publicly available in accordance with Section 734.3(b)(3) and published in accordance with Section 734.7 of the U.S. Export Administration Regulations. As a result of this publication, this report is subject to only copyright protection and does not require any license agreement from EPRI. This notice supersedes the export control restrictions and any proprietary licensed material notices embedded in the document prior to publication.

T&D System Design and Construction for Enhanced Reliability and Power Quality
1010192

Final Report, March 2006

EPRI Project Manager M. Samotyj

ELECTRIC POWER RESEARCH INSTITUTE 3420 Hillview Avenue, Palo Alto, California 94304-1395 PO Box 10412, Palo Alto, California 94303-0813 USA 800.313.3774 650.855.2121 askepri@epri.com www.epri.com

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM: (A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR (B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT. ORGANIZATION THAT PREPARED THIS DOCUMENT EPRI Solutions, Inc.

NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or e-mail askepri@epri.com. Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc. Copyright 2006 Electric Power Research Institute, Inc. All rights reserved.

CITATIONS
This report was prepared by EPRI Solutions, Inc. 942 Corridor Park Blvd. Knoxville, TN 37932 Principal Investigators T. Short D. Crudele J. Harding C. Perry

This report describes research sponsored by the Electric Power Research Institute (EPRI). The report is a corporate document that should be cited in the literature in the following manner: T&D System Design and Construction for Enhanced Reliability and Power Quality. EPRI, Palo Alto, CA: 2006. 1010192.

iii

PRODUCT DESCRIPTION

Transmission and Distribution (T&D) design and construction standards and practices can impact the quality of power supplied to end users. Standards engineers lack clear guidelines on how to optimize the power quality performance of electrical systems. Traditional design and construction standards/practices are based on optimizing cost with system reliability. However, in some cases, improvement in service reliability could come at the expense of service quality. This project combines prior research from within this program and the industry overall to develop a robust resource that consolidates T&D design information and the effects of standards and practices on reliability and quality. The information presented in this report will enable utility engineers to use their assets more cost-effectively and maintain or improve power quality at a lower cost. Results & Findings This project quantifies the power quality impacts of various T&D design standards, providing guidelines and cost/benefit analysis tools for improving quality and reliability through distribution system design, operations, and maintenance practices. A primary goal of this report is to provide engineering guidelines for standards engineers by addressing the power quality implications of design and construction standards/practices from a number of perspectives. Challenges & Objectives This report is intended to provide the necessary background to help a distribution engineer quickly acquire a working knowledge of the T&D system design and construction issues associated with system reliability and power quality. Topics covered include: Basic approaches to power quality and reliability Protection strategies Construction issues Animal protection Construction and maintenance to reduce tree faults Lightning protection, grounding, and arresters Equipment issues Custom configurations Specific power quality issues such as flicker, harmonics, and voltage unbalance

Applications, Values & Use This report provides a detailed look at many aspects of how T&D system design and construction affect power quality and reliability. This report is intended for utility distribution planning engineers, protection engineers, field engineers, and power quality engineers. The report attempts to provide strategies to address many of the most common power quality problems. EPRI Perspective By providing utilities with clear procedures for improving power quality through changes on the T&D system, EPRI is enabling utilities to provide better customer service. Utilities can target changes to specific customers with high power quality needs or can apply changes that improve power quality across the board. By focusing on practical methods, the report provides solutions for many common power quality problems. Approach The project team identified the most common and most important power quality issues and identified several strategies for improving power quality. The team then developed guidelines and calculation procedures for each major strategy. The report includes many examples that illustrate the usage of each procedure and also provides background material and references. Keywords Power Quality Reliability Distribution system System protection Voltage sags Momentary interruptions

vi

ACKNOWLEDGMENTS
Duke Power deserves special thanks for its support of this project. Both Jerry Thompson (now retired) and Lee Taylor provided significant support and advice. Lee Taylor and Duke Power pioneered the strategy of identifying and removing sources of faults on the Duke Power system. By doing this, Duke Power is able to maintain very respectable reliability numbers despite the fact that its service territory has regular sever weather, and Duke Power is mainly an overhead utility with predominantly all-radial systems. Duke Power also provided a number of educational photographs that provide many useful examples.

vii

CONTENTS

1 PROJECT INTRODUCTION ..................................................................................................1-1 Companion Website ..............................................................................................................1-2 2 BASIC APPROACHES ..........................................................................................................2-1 Power Quality Problems........................................................................................................2-1 Problem Identification .......................................................................................................2-1 Solution Options ...............................................................................................................2-2 Voltage Sags ....................................................................................................................2-4 Momentary Interruptions.................................................................................................2-11 Sustained Interruptions...................................................................................................2-12 Reduction of Faults and Targeted Maintenance.............................................................2-14 Using Outage Databases ....................................................................................................2-16 3 PROTECTION STRATEGIES ................................................................................................3-1 Single-Phase Protective Devices ..........................................................................................3-1 Industry Usage .................................................................................................................3-1 Estimating the Improvement.............................................................................................3-2 Single-Phasing of Three-Phase Motors............................................................................3-3 Ferroresonance ................................................................................................................3-4 Backfeed to Faults............................................................................................................3-7 Single-Phase Reclosers With Three-Phase Lockout......................................................3-10 Single-Phase Protective Devices Summary ...................................................................3-11 Instantaneous Reclose........................................................................................................3-12 Benefits ......................................................................................................................3-12 Negatives ...................................................................................................................3-12 Industry Usage ...............................................................................................................3-12 Effect on Sensitive Residential Devices .........................................................................3-13 Delay Necessary to Avoid Retriggering Faults ...............................................................3-14

ix

Reclose Impacts on Motors ............................................................................................3-15 Summary ........................................................................................................................3-16 Fusing .................................................................................................................................3-17 Fuse Saving Versus Fuse Blowing......................................................................................3-18 Fuse Saving...............................................................................................................3-18 Benefits .................................................................................................................3-19 Negatives ..............................................................................................................3-19 Fuse Blowing .............................................................................................................3-19 Benefits .................................................................................................................3-19 Negatives ..............................................................................................................3-19 Economics .................................................................................................................3-20 Bottom Line................................................................................................................3-20 Introduction.....................................................................................................................3-20 Industry Usage ...............................................................................................................3-21 Effects on Momentaries and Sustained Interruptions.....................................................3-22 Coordination Limits of Fuse Saving................................................................................3-23 Long-Duration Faults and Wire Burndowns With Fuse Blowing .....................................3-27 Long-Duration Voltage Sags With Fuse Blowing............................................................3-31 Optimal Implementation of Fuse Saving.........................................................................3-32 Optimal Implementation of Fuse Blowing .......................................................................3-33 Time Delay on the Instantaneous Element (Fuse Blowing)............................................3-34 High-Low Combination Scheme .....................................................................................3-35 Control of the Protection Scheme by SCADA System ...................................................3-36 Adaptive Control by Phases ...........................................................................................3-37 Economics ......................................................................................................................3-37 Summary ........................................................................................................................3-37 Current-Limiting Fuses ........................................................................................................3-38 Benefits ......................................................................................................................3-38 Negatives ...................................................................................................................3-38 Economics .................................................................................................................3-38 Bottom Line................................................................................................................3-38 Introduction.....................................................................................................................3-38 Industry Usage ...............................................................................................................3-40 Types of Current-Limiting Fuses ....................................................................................3-41

Safety Advantages .........................................................................................................3-42 Best Locations for Use ...................................................................................................3-44 Drawbacks......................................................................................................................3-47 Economics ......................................................................................................................3-47 Summary ........................................................................................................................3-47 Sequence Coordination of Reclosers..................................................................................3-47 Benefits ......................................................................................................................3-47 Negatives ...................................................................................................................3-48 Economics .................................................................................................................3-48 Introduction.....................................................................................................................3-48 Standard Coordination and Its Effect on Momentary Interruptions.................................3-48 Sequence Coordination and Its Effect on Momentary Interruptions ...............................3-49 Comparison of Reliability................................................................................................3-50 Coordination Requirements............................................................................................3-52 Summary ........................................................................................................................3-54 Locating Sectionalizing Equipment .....................................................................................3-55 4 CONSTRUCTION ISSUES.....................................................................................................4-1 Fault-Reduction Methods and Processes .............................................................................4-1 Fault Basics ......................................................................................................................4-2 Programs to Improve Construction Practices ...................................................................4-6 Data Analysis and Targeting Programs............................................................................4-7 Field Identification of Fault Sources .................................................................................4-9 Design Review With Power Quality in Mind ...................................................................4-14 Outage Follow-Ups.........................................................................................................4-15 Construction Audits ........................................................................................................4-16 Construction Upgrade Programs ....................................................................................4-18 Problem-Circuit Audits....................................................................................................4-19 High-Tech Designs..............................................................................................................4-23 Industry Practices ...........................................................................................................4-23 Duke Power Case Study ................................................................................................4-24 Underground Systems....................................................................................................4-26 Conductor Spacing and Span Lengths................................................................................4-28 NESC Clearance Requirements.....................................................................................4-29 Covered Conductors and Neutrals .................................................................................4-34

xi

Illustration of Clearances ................................................................................................4-35 Climbing Spaces and Work Areas..................................................................................4-37 Span Lengths .................................................................................................................4-37 Conductor Slapping Due to Short Circuits...........................................................................4-39 Magnetic Forces .............................................................................................................4-41 Possible Conductor Slapping Solutions..........................................................................4-47 Wind and Trees ...................................................................................................................4-47 Tree Faults .....................................................................................................................4-47 Faults Due to Wind-Induced Conductor Motion..............................................................4-49 Mitigating Wind and Tree Faults.....................................................................................4-50 5 ANIMAL PROTECTION .........................................................................................................5-1 Animal-Fault Basics...............................................................................................................5-1 Animal Guards, Wire Coverings, and Other Protective Equipment.......................................5-6 Construction Considerations ...............................................................................................5-15 Ground Wires .................................................................................................................5-15 Support Brackets ............................................................................................................5-17 Fusing.............................................................................................................................5-19 Summary of Animal-Protection Guidelines..........................................................................5-21 6 CONSTRUCTION AND MAINTENANCE PROGRAMS TO REDUCE TREE FAULTS.........6-1 Tree Faults ............................................................................................................................6-1 Momentary Interruptions from Trees............................................................................6-1 Voltage Sags From Trees ............................................................................................6-4 Voltage Flicker From Trees..........................................................................................6-5 Previous Work .............................................................................................................6-9 Tree Fault Case Studies.................................................................................................6-11 Baltimore Gas and Electric ........................................................................................6-11 Eastern Utilities Associates........................................................................................6-11 BC Hydro ...................................................................................................................6-12 Niagara Mohawk ........................................................................................................6-12 Duke Power ...............................................................................................................6-12 Environmental Consultants Inc. .................................................................................6-13 Southeastern US Utility..............................................................................................6-13 Characteristics of Tree Faults and Interruptions.............................................................6-14

xii

Physics of Tree Faults ....................................................................................................6-23 Covered Conductors and Other Construction Approaches to Reduce Tree Faults ............6-26 Underground Circuits......................................................................................................6-26 Circuit Location...............................................................................................................6-26 Covered Conductors, Spacer Cable, and Aerial Cable ..................................................6-26 Industry Performance Data.............................................................................................6-30 Covered Wire Issues ......................................................................................................6-32 Arc Damage and Burndowns .....................................................................................6-32 Controlling Arc Damage.............................................................................................6-36 Arc Protective Devices...............................................................................................6-37 Lightning Protection and Covered Conductors ..........................................................6-39 Wire Tie and Insulator Compatibility ..........................................................................6-40 Other Covered-Conductor Issues ..............................................................................6-42 Mechanical Coordination of Construction.......................................................................6-43 Utility Programs to Reduce Tree-Related Power Quality Problems ....................................6-45 Tree Maintenance...........................................................................................................6-45 General Guidelines.........................................................................................................6-45 Program Cost .................................................................................................................6-48 Tree-Trimming Cycle ......................................................................................................6-51 Hazard-Tree Programs...................................................................................................6-55 Identifying Tree Defects..................................................................................................6-57 Right-of-Way Widening...................................................................................................6-64 Audits..............................................................................................................................6-64 Utility Results With Targeted Programs..........................................................................6-64 Eastern Utilities ..........................................................................................................6-64 Niagara Mohawk ........................................................................................................6-65 Puget Sound Energy..................................................................................................6-66 Other Utility Programs................................................................................................6-66 7 LIGHTNING PROTECTION, GROUNDING, AND ARRESTERS ..........................................7-1 Lightning Protection Background ..........................................................................................7-1 Direct Strikes ....................................................................................................................7-2 Induced Voltages..............................................................................................................7-4 Surge Arresters and Arrester Reliability ................................................................................7-7 Older Arresters ...............................................................................................................7-10

xiii

Arrester Installation Considerations................................................................................7-12 Pole Insulation and Lightning ..............................................................................................7-14 Insulation and Pole Construction Guidelines..................................................................7-15 Line Protection ....................................................................................................................7-22 Shield Wires ...................................................................................................................7-22 Line Protection Arresters ................................................................................................7-24 Grounding ...........................................................................................................................7-26 Secondary-Side Surges and Grounding.........................................................................7-28 Summary of Guidelines for Lightning Protection, Grounding, and Arrester Application......7-30 8 EQUIPMENT ISSUES ............................................................................................................8-1 Fiberglass Standoffs..............................................................................................................8-1 Mechanics of the Flash Over............................................................................................8-5 Are Flashovers More Likely After Thunderstorms? ..........................................................8-8 Laboratory Demonstration ................................................................................................8-9 Degradation of Other Fiberglass Apparatus .....................................................................8-9 Possible Solutions ..........................................................................................................8-10 The Need for Further Study............................................................................................8-10 Polymer Insulators...............................................................................................................8-11 Composition of Polymer Insulators.................................................................................8-11 Available Designs ...........................................................................................................8-12 Advantages of Polymer Insulation ..................................................................................8-13 Drawbacks of Polymer Insulation ...................................................................................8-13 Degradation of Polymer Insulators .................................................................................8-14 Issues Not Related to Electrical Performance ................................................................8-15 Current Usage Trends ....................................................................................................8-15 Stress Cracking of Fuse Cutouts.........................................................................................8-15 Porcelain Cutout Construction and Crack Formation .....................................................8-16 Approaches to Risk Management for Porcelain Cutout Cracking ..................................8-17 9 CUSTOM CONFIGURATIONS...............................................................................................9-1 Reactors ................................................................................................................................9-1 Benefits ........................................................................................................................9-1 Negatives .....................................................................................................................9-1 Economics ...................................................................................................................9-1

xiv

Bottom Line..................................................................................................................9-1 Introduction.......................................................................................................................9-2 Available Short-Circuit Current: Traditional Substation Design ........................................9-2 Drop-In Substations..........................................................................................................9-3 Distribution System Sag Performance..............................................................................9-3 Nature of Distribution System Events...............................................................................9-4 Excessive Fault Duties .....................................................................................................9-5 Reactor Placement Options .........................................................................................9-5 Power Quality Ramification ..............................................................................................9-9 Study Findings................................................................................................................9-10 Station Neutral Reactors and Customer Transformer Connections ...............................9-11 Conclusion......................................................................................................................9-13 Advanced Customer-Service Configurations ......................................................................9-13 Customer-Specific Solutions......................................................................................9-13 Benefits ......................................................................................................................9-14 Negatives ...................................................................................................................9-14 Economics .................................................................................................................9-14 Bottom Line................................................................................................................9-14 Reliability Versus Availability ..........................................................................................9-14 Examples of Critical Installations....................................................................................9-15 Customer Perception......................................................................................................9-16 Interruption and Sag Frequency .....................................................................................9-17 Whole-Site Mitigation......................................................................................................9-18 Primary-Side Transfer Switches .....................................................................................9-19 Standard Transfer Switch Concerns ..........................................................................9-21 Load Faults ...........................................................................................................9-21 Delayed Transfer...................................................................................................9-21 Motor Transfer.......................................................................................................9-21 Mechanical Wear ..................................................................................................9-24 Potential Changes to Existing Operation ...................................................................9-24 Ultra-Fast Primary-Side Transfer Switches ....................................................................9-24 FastBut Not SubcycleTransfer Switches .................................................................9-27 Secondary-Side Transfer Switches ................................................................................9-30 Custom Power Equipment..............................................................................................9-30

xv

Spot Networks ................................................................................................................9-31 Summary ........................................................................................................................9-33 10 SPECIFIC POWER QUALITY ISSUES ..............................................................................10-1 Voltage Regulation ..............................................................................................................10-1 Solutions to Voltage-Regulation Problems .....................................................................10-2 Voltage Unbalance ..............................................................................................................10-3 Solutions to Voltage Unbalance .....................................................................................10-6 Voltage Flicker ....................................................................................................................10-7 Motor Starting .................................................................................................................10-8 Solutions to Voltage Flicker ............................................................................................10-9 Capacitor Switching and Tripping of Adjustable-Speed Drives .........................................10-10 Effect on Customer Equipment.....................................................................................10-12 Solutions to Switching Transients.................................................................................10-14 Harmonics .........................................................................................................................10-16 Solutions to Harmonics.................................................................................................10-19 Telephone Interference .....................................................................................................10-20 Solutions to Telephone Interference.............................................................................10-21 11 SUMMARY .........................................................................................................................11-1

xvi

LIST OF FIGURES
Figure 1-1 Screenshot from www.pqtd.com...............................................................................1-2 Figure 2-1 Example Distribution System Showing Fault Location and Customer Impact ..........2-5 Figure 2-2 Example Voltage Sag During a Fault .......................................................................2-5 Figure 2-3 Voltage Divider Illustration Giving the Bus Voltage for a Downstream Fault ............2-7 Figure 2-4 Substation Voltage Profile for Faults at the Given Distance (Single-Phase and Three-Phase Faults Are Shown for Each Voltage) for a 25-MVA, 10% Transformer and 350-kcmil Feeder ........................................................................................................2-9 Figure 2-5 Cumulative Distribution of Substation Bus Voltage Sags Per Year for the Given (25-MVA, 10% Transformer, 350-kcmil Feeder, n = 2 or 4 Feeders off the Bus, =1 Faults/Phase/Mile of Mains/Year, Assumes Line-to-Ground Faults Only) ........2-10 Figure 2-6 One-Line Diagram Showing the Boundaries Where Faults Will Cause Sags, Momentaries, and Interruptions for the Customer Shown................................................2-16 Figure 3-1 Derating Curve for Induction/Synchronous Motors (NEMA MG-1 Part 20 and 21) ......................................................................................................................................3-4 Figure 3-2 Example of Ferroresonance Measured During a Manual, Single-Pole Switching Operation as Part of the EPRI Distribution Power Quality Project.....................3-5 Figure 3-3 Ferroresonance on a Cable-Fed Transformer With an Ungrounded High-Side Connection .........................................................................................................................3-6 Figure 3-4 Equivalent Circuits of Ferroresonance on a Transformer With an Ungrounded High-Side Connection ........................................................................................................3-6 Figure 3-5 Backfeed to a Downed Conductor............................................................................3-8 Figure 3-6 Available Backfeed Current on a 12.47-kV Circuit (Wye-Wye- or Delta-WyeConnected Transformer) ....................................................................................................3-9 Figure 3-7 Available Backfeed Voltage on a 12.47-kV Circuit .................................................3-10 Figure 3-8 Unfused CSP on a Mainline ...................................................................................3-18 Figure 3-9 Sample Circuit for Use in Comparing Fuse Saving and Fuse Blowing During a Fault on a Lateral .............................................................................................................3-20 Figure 3-10 Comparison of Fuse Saving and Fuse Blowing on Hypothetical Circuit...............3-23 Figure 3-11 Coordination of a 100-K Lateral Fuse With a 5-Cycle Breaker.............................3-24 Figure 3-12 Burndown Threshold of Damage Compared to Typical Relay Characteristics.....3-29 Figure 3-13 Bare-Wire Burndown Curves for All-Aluminum Alloy Conductor Along With a 100-K Lateral Fuse and a Ground Relay Element ...........................................................3-31 Figure 3-14 Magnitudes and Durations of Substation Bus Voltage Sags for Ground Faults Applied Every Half Mile on a 12.47-kV Circuit With a Fuse-Blowing Scheme.......3-32 Figure 3-15 Example of a Delayed Instantaneous Element Used for Fuse Blowing................3-35

xvii

Figure 3-16 Voltage and Current Measured During Operation of a Current-Limiting Fuse in the EPRI Distribution Power Quality Study...................................................................3-39 Figure 3-17 Voltage Sags on the Station Bus Versus Fault Location as a Function of Distance From the Station................................................................................................3-45 Figure 3-18 Fault Current Versus Fault Location as a Function of Distance From the Station ..............................................................................................................................3-46 Figure 3-19 Performance of Reclosers With and Without Sequence Coordination .................3-49 Figure 3-20 Example Distribution Circuit .................................................................................3-50 Figure 3-21 Coordination of the Fast Curves of Two Reclosers ..............................................3-53 Figure 3-22 Coordination of the Fast Curves of Two Reclosers, One With Minimum Response Set to 0.06 Seconds........................................................................................3-54 Figure 3-23 Recloser Examples with Customer Interruptions Saved.......................................3-56 Figure 3-24 Tap Fuse Example with Customer Interruptions Saved .......................................3-57 Figure 3-25 Duke Power Sectionalizing Example ....................................................................3-58 Figure 4-1 Fault Causes Measured in the EPRI Fault Study .....................................................4-2 Figure 4-2 Fault Rates Found in Different Studies.....................................................................4-4 Figure 4-3 Power Arc Damage on the Base of a Fiberglass Insulator Standoff.........................4-6 Figure 4-4 Example Distribution System Showing Fault Locations and Their Impact on One Customer....................................................................................................................4-8 Figure 4-5 One-Line Diagram Showing the Boundaries Where Faults Will Cause Sags, Momentaries, and Interruptions for the Customer Shown..................................................4-9 Figure 4-6 Fault Arc Damage on a Gapped Arrester ...............................................................4-11 Figure 4-7 Transformer Pole With Minimal Phase-to-Phase Clearances and No Animal Guards .............................................................................................................................4-12 Figure 4-8 Poor Clearances and Evidence of Arcing Across an Insulator ...............................4-13 Figure 4-9 Blown Arrester ........................................................................................................4-14 Figure 4-10 Deteriorated Pole Top ..........................................................................................4-20 Figure 4-11 Riser-Pole Installation With Evidence of Multiple Arc Burns.................................4-20 Figure 4-12 Trees Growing Into Lines .....................................................................................4-21 Figure 4-13 Lines With Significant Tree Overhang ..................................................................4-21 Figure 4-14 High-Reliability Design Example ..........................................................................4-24 Figure 4-15 Duke Power Outage-Resistant Design .................................................................4-25 Figure 4-16 Duke Power Double-Circuit Outage-Resistant Design .........................................4-26 Figure 4-17 Application of Horizontal Clearances at Supports Between Conductors and Vertical Clearance Between a Road and Conductors ......................................................4-35 Figure 4-18 Vertical Clearances Between Conductors ............................................................4-36 Figure 4-19 Geometries Involving Both Horizontal and Vertical Clearances ...........................4-36 Figure 4-20 Example of a Climbing Space That Has Been Provided in the Layout of the Pole to Allow Safe Crew Access ......................................................................................4-37 Figure 4-21 Illustration of Conductor Slapping Fault Due to Magnetic Forces of an Earlier Flashover .........................................................................................................................4-40

xviii

Figure 4-22 A Sagging Conductor Can Be Represented by a Mass Equal to the Weight of One Span With an Arm Length Equal to 2/3 of the Total Sag of the Conductor ..........4-43 Figure 4-23 Theoretical Illustration of the Conductor Displacement Versus Time ...................4-44 Figure 4-24 Shows the Method of Conservation of Energy to Calculate the Height H of the Initial Swing ................................................................................................................4-46 Figure 4-25 Example of a Non-Conventional Design That Might be More Resistant to Certain Tree Faults ..........................................................................................................4-49 Figure 5-1 A Squirrel Flirting With Danger .................................................................................5-2 Figure 5-2 Ranges of Squirrel Species ......................................................................................5-3 Figure 5-3 Nest in a Transformer...............................................................................................5-4 Figure 5-4 Overhead Distribution Points Most Susceptible to Animal-Caused Faults ...............5-6 Figure 5-5 Various Animal Guards.............................................................................................5-7 Figure 5-6 Examples of Installations With Animal Guards and Insulated Jumpers....................5-8 Figure 5-7 Animal Guard Installation .......................................................................................5-10 Figure 5-8 Examples of Poor Animal Guard Installations ........................................................5-10 Figure 5-9 Problems Noted at Poles With Missing or Improperly Installed Animal Protective Devices ...........................................................................................................5-11 Figure 5-10 Example of a Deteriorated Animal Guard Compared to a New One ....................5-13 Figure 5-11 Electrocuted Squirrel and the Ground Wire That Made the Pole Susceptible to Animal Faults ...............................................................................................................5-16 Figure 5-12 Example of a Grounded Bracket ..........................................................................5-18 Figure 5-13 Examples of Insulated Supports...........................................................................5-19 Figure 5-14 CSP Transformer Exposing an Entire Circuit Tap to Animal Faults .....................5-20 Figure 5-15 Line Arresters That Leave the Mainline Exposed to Animal Faults Across the Arresters...........................................................................................................................5-21 Figure 5-16 Pole Charred After a Nest Fire .............................................................................5-22 Figure 6-1 Feeder Lockouts Versus Momentary Events............................................................6-3 Figure 6-2 Feeder Lockouts Versus Momentary Events by Month ............................................6-4 Figure 6-3 GE Voltage Flicker Curve .........................................................................................6-5 Figure 6-4 Resistance Measurements on a Live Poplar Tree....................................................6-7 Figure 6-5 Tree Resistivities ......................................................................................................6-8 Figure 6-6 Niagara Mohawk Survey of Tree Outage Causes ..................................................6-12 Figure 6-7 ECI Survey of Tree Outage Causes .......................................................................6-13 Figure 6-8 Breakdown of Reliability Indices for a Northeastern Utility .....................................6-14 Figure 6-9 Percentage of Interruptions by Cause for the Given Weather for a Northeastern Utility...........................................................................................................6-15 Figure 6-10 Likelihood of the Given Cause for the Specified Weather for a Northeastern Utility ................................................................................................................................6-16 Figure 6-11 Interruption Indices by Cause During Major Storms and With and Without Major Storms....................................................................................................................6-17 Figure 6-12 Tree-Caused Outage Impacts by Month ..............................................................6-18

xix

Figure 6-13 Interruptions by Month for One Northeastern Utility (Major Events Included) ......6-19 Figure 6-14 Tree-Caused Customer Interruptions by Interrupting Device for One Northeastern Utility (Major Events Included)....................................................................6-20 Figure 6-15 Allocation of Tree-Caused System SAIDI by Feeder for One Northeastern US Utility ..........................................................................................................................6-21 Figure 6-16 Allocation of Tree-Caused System SAIDI by Tree Outage for One Northeastern US Utility.....................................................................................................6-21 Figure 6-17 Percentage of Samples Faulted Based on the Voltage Gradient Across the Tree Branch .....................................................................................................................6-24 Figure 6-18 Time to Fault Based on the Voltage Gradient Across the Tree Branch................6-25 Figure 6-19 Example of a Compact Armless Design Using Covered Conductors ...................6-28 Figure 6-20 Example of a Spacer Cable Run Through Trees..................................................6-28 Figure 6-21 Conductor Damage From Arcing Where the Conductor Cover Begins ................6-33 Figure 6-22 Conductor Damage From Arcing..........................................................................6-33 Figure 6-23 Burndown Characteristics of Various Conductors ................................................6-36 Figure 6-24 Arc Protective Devices .........................................................................................6-38 Figure 6-25 An Arc Protective Device That Has Operated ....................................................6-39 Figure 6-26 Example of a Covered Wire Tie on a Covered Conductor ...................................6-41 Figure 6-27 Example of Damage on a Covered Conductor From Flashover at the Insulator Tie .....................................................................................................................6-42 Figure 6-28 A Pole Broken in Half by a Tree Falling Onto the Line Structure During a Windstorm ........................................................................................................................6-44 Figure 6-29 Circuits With Significant Tree Overhang...............................................................6-46 Figure 6-30 Circuits With Impending Tree Growth Contact .....................................................6-47 Figure 6-31 Circuits With Vines ...............................................................................................6-48 Figure 6-32 Utility Vegetation Management Costs of Five Utilities Surveyed ..........................6-49 Figure 6-33 Costs of Vegetation Management From Several Utilities .....................................6-50 Figure 6-34 Vegetation Management Costs Versus Performance...........................................6-51 Figure 6-35 Tree Maintenance Effect on SAIFI and SAIDI for a Northeastern Utility ..............6-52 Figure 6-36 Tree Maintenance Effect on Tree Outage Rate Per Mile......................................6-52 Figure 6-37 Tree Maintenance Effect on Tree Outage Rate Per Mile for a Southeastern Utility ................................................................................................................................6-53 Figure 6-38 Defects Causing Tree Failure for the Niagara Mohawk Power Corporation.........6-56 Figure 6-39 Example of Deadwood .........................................................................................6-58 Figure 6-40 Examples of Cracks..............................................................................................6-59 Figure 6-41 Example of a Strong Branch Union ......................................................................6-60 Figure 6-42 Example of a Weak Branch Union........................................................................6-60 Figure 6-43 Examples of Outward Signs of Tree Decay Discoloration, Holes, and Fungal Activity..................................................................................................................6-61 Figure 6-44 Examples of Cankers ...........................................................................................6-61

xx

Figure 6-45 Example of Root Damage due to Excavation.......................................................6-62 Figure 6-46 Example of Crown Decline due to Root Damage .................................................6-62 Figure 6-47 Examples of Leaning Trees due to Root Damage................................................6-62 Figure 6-48 Example of Poor Tree Architecture ......................................................................6-63 Figure 7-1 Distribution Line Fault Rate Versus Ground Flash Density ......................................7-2 Figure 7-2 Induced Voltage Flashovers Versus Insulation Level for a Line With a Grounded Neutral...............................................................................................................7-7 Figure 7-3 Examples of Older Arresters ..................................................................................7-11 Figure 7-4 Example of a Gapped Arrester on a Recloser........................................................7-12 Figure 7-5 Blown Arrester With a Dangling Ground Lead........................................................7-13 Figure 7-6 Example of a Blown Arrester ..................................................................................7-14 Figure 7-7 Example of an Armless Design ..............................................................................7-16 Figure 7-8 Example Pole-Protection Assemblies.....................................................................7-17 Figure 7-9 Example of Wood Poles Bonded With Ground Wires ............................................7-18 Figure 7-10 Examples of Poles With Probable Lightning Damage ..........................................7-19 Figure 7-11 Example of Uninsulated Guy Wires Near Phase Conductors...............................7-20 Figure 7-12 Example of a Poorly Insulated Guy Wire Near Phase Conductors.......................7-21 Figure 7-13 A Fiberglass Guy Insulator Shorted Out ...............................................................7-21 Figure 7-14 Shield-Wire Lightning Protection System .............................................................7-23 Figure 7-15 Performance of a Shield Wire Depending on Grounding and Insulation Level.....7-23 Figure 7-16 Shield Wire Shielding Angle .................................................................................7-24 Figure 7-17 Impact of Grounding .............................................................................................7-27 Figure 7-18 Lightning Surge Entry via the Secondary .............................................................7-29 Figure 8-1 Wooden Cross Arm Construction .............................................................................8-2 Figure 8-2 Fiberglass Standoff Construction .............................................................................8-3 Figure 8-3 A Weathered and Damaged Fiberglass Standoff Showing the Various Components.......................................................................................................................8-4 Figure 8-4 Damaged Fiberglass Standoff ..................................................................................8-4 Figure 8-5 Close-up View of Damaged Fiberglass Standoff......................................................8-5 Figure 8-6 Vertically Mounted Fiberglass Standoff Bracket Showing Surface Degradation ......8-6 Figure 8-7 Horizontally Mounted Fiberglass Standoff Bracket Showing Surface Degradation........................................................................................................................8-6 Figure 8-8 Fiberglass Standoff Bracket Showing Surface Degradation.....................................8-7 Figure 8-9 Capacitive Voltage Divider Created Between a Ceramic Pin-Type Insulator and the Fiberglass Standoff ...............................................................................................8-8 Figure 8-10 Discharges Over the Surface of an Energized Fiberglass Standoff .......................8-9 Figure 8-11 Different Polymer Insulator Designs .....................................................................8-13 Figure 8-12 Typical Porcelain Fuse Cutout Configuration Showing Cemented Metal Pins in Porcelain ......................................................................................................................8-16 Figure 9-1 Fault Duty on Single Transformer for Line Fault.......................................................9-2

xxi

Figure 9-2 Fault Duty for Two Transformers for a Line Fault .....................................................9-3 Figure 9-3 Data from EPRI Distribution Power Quality Study Overlaid on the ITIC (Information Technology Industry Council) Curve ..............................................................9-4 Figure 9-4 Fault Duty With Tie Reactor .....................................................................................9-6 Figure 9-5 Effect of Transformer Secondary Reactors on Fault Duty ........................................9-7 Figure 9-6 Effect of Feeder Reactors on Fault Duty ..................................................................9-8 Figure 9-7 Common Distribution Station Arrangement ..............................................................9-9 Figure 9-8 Bus Voltage (in Per Unit) During Line-to-Ground Faults as a Function of Distance From the 23-kV Bus to the Fault .......................................................................9-10 Figure 9-9 Percentage of Total Circuit Length Relative to Distance From Bus........................9-11 Figure 9-10 Neutral Reactor on the Substation Transformer Along With Delta-Wye Transformers....................................................................................................................9-12 Figure 9-11 Sag and Interruption Rate From EPRI Distribution Power Quality Study .............9-17 Figure 9-12 Distribution of Sag Durations From a National Power Laboratory Study..............9-18 Figure 9-13 One-Line Diagram of a Dual-Feed Transfer Scheme ...........................................9-19 Figure 9-14 One-Line Diagram of a Dual-Feed Transfer Scheme With Backup Generation .......................................................................................................................9-20 Figure 9-15 Motor Vector in Relation to Fixed Utility Vector During an Interruption ................9-22 Figure 9-16 Motor Terminal Voltages at Start of a Power Outage ...........................................9-23 Figure 9-17 Motor Terminal Voltage at 6 Seconds After the Start of a Power Outage ............9-24 Figure 9-18 Typical Static Switch Configuration ......................................................................9-25 Figure 9-19 Voltage on Two Feeders Served From a Common Bus With a Fault on One Feeder..............................................................................................................................9-26 Figure 9-20 Normal Voltage Phasors for a Three-Phase Service............................................9-28 Figure 9-21 Voltage Phasors for a Single-Line-to-Ground Fault on Phase C ..........................9-29 Figure 9-22 A Spot Secondary Network ..................................................................................9-32 Figure 9-23 Comparison of a Traditional Spot Network to a Configuration With Added Reactors to Minimize the Voltage Sags Caused by Primary Feeder Faults.....................9-33 Figure 10-1 ANSI C84.1 Voltage Ranges ................................................................................10-2 Figure 10-2 Approximate Percent Voltage Unbalance on the U.S. Distribution System..........10-5 Figure 10-3 Fuse Blown on a Capacitor BankA Common Cause of Voltage Unbalance .....10-6 Figure 10-4 GE Flicker Curve ..................................................................................................10-7 Figure 10-5 Transient Caused by Switching of Capacitor at a Random Point of the Sine Wave ..............................................................................................................................10-11 Figure 10-6 Illustration of a Restrike on a Grounded-Wye Bank ...........................................10-12 Figure 10-7 Effect of Capacitor-Switching Transient on the Direct Current Bus of an Adjustable-Speed Drive .................................................................................................10-13 Figure 10-8 Transient Caused by Synchronous Switching of a Capacitor Bank....................10-15 Figure 10-9 Waveform and Harmonic Spectrum of Typical 6-Pulse ac Motor Drive..............10-17 Figure 10-10 Harmonic Resonance .......................................................................................10-18

xxii

Figure 10-11 Tuned Harmonic Filter ......................................................................................10-19 Figure 10-12 Telephone Influence Factor (TIF) Curve ..........................................................10-21

xxiii

LIST OF TABLES
Table 2-1 Summary of Utility-Side Options for Improving Power Quality Within the Distribution System ............................................................................................................2-3 Table 2-2 Customer-Side Options for Improving Power Quality ................................................2-4 Table 2-3 Cumulative Voltage Sag Table for NPL Data With 5-Minute Filter: Number of Sags per Year ....................................................................................................................2-6 Table 2-4 Typical Line Impedances (Ohms/Mile) ......................................................................2-7 Table 2-5 Typical Transformer Impedances (Z1 = Z0 = Zground fault = jZ%*kV /MVA) Ohms With Bolted-Bus Fault Currents Given in Parenthesis........................................................2-8 Table 2-6 Average Number of Momentary Interruptions Found in Several Studies.................2-11 Table 2-7 Reliability Data From a 1995 Survey by the Institute of Electrical and Electronics Engineers ......................................................................................................2-14 Table 2-8 Reliability Data With and Without Storms from a 1998 Survey by Edison Electric Institute (EEI).......................................................................................................2-14 Table 3-1 Effect on Interruptions When Using Single-Phase Protective Devices on Three-Phase Circuits .........................................................................................................3-2 Table 3-2 Line-to-Ground and Line-to-Line Voltages With One Phase Interrupted ...................3-3 Table 3-3 Transformer Primary Connections Susceptible to Ferroresonance ...........................3-6 Table 3-4 Highest Current Ratings of Single-Phase Reclosers That Have Three-Phase Lockout Capability or Three-Phase Reclosers With Single-Phase Trip ...........................3-11 Table 3-5 Percent of Distribution Lines Using Specified Intervals Between Reclose Attempts Based on an IEEE Survey ................................................................................3-13 Table 3-6 Percent of Devices That Were Able to Successfully Ride Through a Momentary Interruption of Specified Durations ................................................................3-13 Table 3-7 A Comparison of the Sequence of Events for Fuse Saving and Fuse Blowing .......3-21 Table 3-8 Results of Surveys by the Institute of Electrical and Electronics Engineers on the Percentage of Utilities That Use Fuse Saving............................................................3-21 Table 3-9 1996 PTI Survey on Line-Protection Methods .........................................................3-22 Table 3-10 Maximum Fault Currents for Fuse Saving Coordination for Several Common Fuse Links for a 5-Cycle Breaker and a 1-Cycle Relay Time...........................................3-25 Table 3-11 Maximum Fault Currents for Fuse Saving Coordination for Several Common Fuse Links for a 3-Cycle Breaker and a -Cycle Relay Time..........................................3-26 Table 3-12 Minimum Distance From the Substation Where the Given Fuse Coordinates for Fuse Saving for a 5-Cycle Breaker and a 1-Cycle Relay Time (12.47 kV, 500 kcmil Phase, 4/0 Neutral) (For 24.94 kV, Multiply Distances by 2) ..................................3-26
2

xxv

Table 3-13 Minimum Distance From the Substation Where the Given Fuse Coordinates for Fuse Saving for a 3-Cycle Breaker and a -Cycle Relay Time (12.47 kV, 500 kcmil Phase, 4/0 Neutral) (For 24.94 kV, Multiply Distances by 2) ..................................3-27 Table 3-14 Percent of Utilities Using Different Types of Current-Limiting Fuses as Reported in a 1995 Institute of Electrical and Electronics Engineers Survey...................3-41 Table 3-15 Maximum - to 1-Cycle Current Rating on Distribution Transformers ..................3-43 Table 3-16 Maximum I t Values for Common Current-Limiting Fuses .....................................3-44 Table 3-17 Typical Transformer I t Capabilities .......................................................................3-44 Table 3-18 Number of Device Operations for a Fault in the Specified Zone (Assuming Two Fast and Two Delayed Operations Before Lockout) ................................................3-51 Table 3-19 Duke Power Sectionalizing Cost Guidelines..........................................................3-58 Table 4-1 Number of Phases Involved in Each Fault Measured in the EPRI Fault Study .........4-3 Table 4-2 Normal Temporary or Permanent Fault Causes ........................................................4-5 Table 4-3 Performance and Cost Comparisons of Duke Powers Double-Circuit OutageResistant Design ..............................................................................................................4-26 Table 4-4 Overhead Versus Underground: Advantages of Each.............................................4-27 Table 4-5 Performance and Cost Comparisons.......................................................................4-28 Table 4-6 Horizontal Clearances Between Line Conductors Smaller Than AWG No. 2 at Supports, Based on Sags (Adopted From NESC C2-2002 Table 235-2) ........................4-31 Table 4-7 Horizontal Clearances Between Line Conductors AWG No. 2 or Larger at Supports, Based on Sags (Adopted From NESC C2-2002 Table 235-3) ........................4-32 Table 4-8 Vertical Clearances Between Conductors at Supports ............................................4-33 Table 4-9 Vertical Clearance Requirements Above Ground, Roadways, Rail Track and Water................................................................................................................................4-34 Table 4-10 Force Per Linear Foot of Conductor for a Line-to-Line Fault on a Feeder.............4-42 Table 4-11 Short Circuit Forces for Typical Constructions.......................................................4-43 Table 4-12 Period of Conductor Oscillation for Various Sags (Assuming no Damping) ..........4-45 Table 5-1 Surveyed Use of Animal-Mitigating Measures on Overhead Distribution ..................5-9 Table 5-2 Deterioration of Animal-Protective Devices .............................................................5-14 Table 6-1 Current Draw Necessary to Cause Observable Light Flicker ....................................6-6 Table 6-2 Current Drawn by Several Trees in Contact With 7.62 kV From Line to Ground.......6-7 Table 6-3 Tree Branch Resistance ............................................................................................6-9 Table 6-4 Percentage of Tree Faults in Each Category...........................................................6-14 Table 6-5 Tree Faults by Voltage Class for a Northeastern US Utility .....................................6-22 Table 6-6 Tree Faults by Voltage Class for a Southeastern US Utility ....................................6-22 Table 6-7 Typical Covering Thicknesses of Covered All-Aluminum Conductor.......................6-27 Table 6-8 Interruption Rates for Various Constructions for CL&P ...........................................6-30 Table 6-9 Comparison of the Reliability Index SAIDI...............................................................6-31 Table 6-10 Burndown Characteristics of Various Conductors .................................................6-35 Table 6-11 Illinois Utility Tree Maintenance Cycle Targets ......................................................6-55
2 2

xxvi

Table 6-12 Comparison of Trees Causing Permanent Faults With the Tree Population .........6-57 Table 6-13 Results of the PSE TreeWatch Program ...............................................................6-66 Table 7-1 Comparison of Induced Voltage Measurements to Rusck Predictions ......................7-6 Table 7-2 Approximate Energy and Charge in ANSI/IEEE Arrester Test Current Waves .........7-9 Table 7-3 CFO of Common Distribution Components (by Themselves)..................................7-15 Table 9-1 Line-to-Ground and Line-to-Line Voltages on the Low-Voltage Side of a Transformer With One Phase on the High-Voltage Side Sagged to Zero........................9-13 Table 9-2 Critical Service Loss Duration for Industrial Plants ..................................................9-16 Table 9-3 Critical Service Loss Duration for Commercial Buildings.........................................9-16 Table 9-4 Ultra-Fast Primary-Side Transfer Switches .............................................................9-25 Table 9-5 Manufacturers of High-Speed, Medium-Voltage, Mechanical Transfer Switches...........................................................................................................................9-27 Table 9-6 Custom Power Device Application Matrix ................................................................9-31 Table 9-7 Comparison of Customer-Specific Configurations ...................................................9-34 Table 11-1 Summary of Utility-Side Options for Improving Power Quality Within the Distribution System ..........................................................................................................11-2

xxvii

1
PROJECT INTRODUCTION
T&D design and construction standards and practices can impact the quality of power supplied to end users. Standards engineers lack clear guidelines on how to optimize the power quality performance of electrical systems. Traditional design and construction standards/practices are based on optimizing cost with system reliability. However, in some cases, improvement in service reliability could come at the expense of service quality. This project quantifies the power quality impacts of various T&D design standards, providing guidelines and cost/benefit analysis tools for improving quality and reliability through distribution system design, operations, and maintenance practices. A primary goal of this report is to provide engineering guidelines for standards engineers, addressing the power quality implications of design and construction standards/practices from a number of perspectives: Line lightning protection Distribution transformer configuration Line clearance Protective device coordination Recloser and fuse operation Capacitor placement and control Single-phase tripping

This project series has spanned several years. Previous work in the series addressed design 1,2 factors that have the most impact on voltage sags, momentary interruptions, and reliability . All are caused by faults on the power system, mostly at the distribution level. The research evaluated practices and equipment to minimize fault impacts on customers and improve system performance. Project funders identified priority design and maintenance practices to be considered, including fault reduction methods and processes, high-tech designs, animal protection, lightning protection, protective device coordination (including relay settings and schemes), tree trimming, underground issues, and looped system configurations and protection. This years project combines prior research, both within this program and within the industry overall, to develop a robust resource that consolidates T&D design information and the effects on reliability and quality. The information presented in this report enables utility engineers to
1 2

Power Quality Improvement Methodology for Wires Companies, EPRI, Palo Alto, CA: 2003. 1001665. Power Quality Implications of Transmission and Distribution Construction: Tree Faults and Equipment Issues, EPRI, Palo Alto, CA: 2005. 1008506.

1-1

Project Introduction

more cost-effectively strategize their asset utilization, and maintain or improve power quality at a lower cost.

Companion Website
There is a companion website to this report that can be found at www.pqtd.com. Log-in is required to access the websites content and it is only available to those funding the project at the time of publication of this report. The website contains articles related to the various topics covered in this report, as well as several on-line calculators and links to EPRI resources and other pertinent information. Figure 1-1 shows a typical screenshot from the companion website.

Figure 1-1 Screenshot from www.pqtd.com

1-2

2
BASIC APPROACHES
Power Quality Problems
The three most-significant power quality concerns for most customers are: Voltage sags Momentary interruptions Sustained interruptions

Power quality problems affect different customers differently. Most residential customers are affected by sustained and momentary interruptions (also known as momentaries). For commercial and industrial customers, sags and momentaries are the most common problems. Each circuit is different, and each customer responds differently to power quality disturbances. The three power quality problems discussed here are caused by faults on the utility power system, with most of them on the distribution system. This report concentrates on practices and equipment that can be used to minimize the effect of faults on customers. Faults can never be totally eliminated, but several ways exist to minimize the impact on customers. Of course, several other types of power quality problems can occur, but these three are the most common. Problem Identification The lights are blinking is the most common customer complaint to utilities. Other common complaints are flickering, clocks blinking, and power out. The first step to improving power quality is identifying what actually is the problem. Sustained interruptions are the easiest to classify because the power is usually out when the customer calls. The lights blinking is harder to classify: Is it momentaries caused by faults on the feeder? Is it voltage sags caused by faults on lateral taps or adjacent feeders? Is it periodic voltage flicker caused by an arc welder or some other fluctuating load on the same circuit? Some strategies for identifying the problem are:

2-1

Basic Approaches

For commercial or industrial customers, does the customer lose all computers or just some of them? Losing all computers probably would indicate that the problem is caused by momentaries. Losing some would indicate that the problem is caused by voltage sags. Is it just the lights flickering? Do any computers or other electronic equipment reboot or reset? If it is just the lights, then the problem is likely to be voltage flicker caused by some fluctuating load (which could be in the facility that is complaining). If approximate times of events are available from the customer, then these times can be compared against the times when utility protective devices operate. Of course, to do this, the utility times must be recorded by a supervisory control and data-acquisition (SCADA) system or a digital relay or recloser controller. If these times are available, it is often possible to correlate a customer outage to the operation of a utility protective device. If the protective device is on an adjacent circuit or the subtransmission system, then the cause was a voltage sag. A review of the number of operations of the protective devices on the circuit (if these records are kept) can reveal whether the customer is seeing an abnormal number of momentaries or possibly sags from faults on adjacent feeders. Are other customers on the circuit having problems? If so, then the problem is probably attributable to momentaries and not just a customer that is extremely sensitive to sags.

Solution Options This report covers many options and technologies for improving the power quality on the utility side. Table 2-1 shows a summary of the techniques discussed in this report.

2-2

Basic Approaches

Table 2-1 Summary of Utility-Side Options for Improving Power Quality Within the Distribution System Voltage Sags

Use fuse saving Use current-limiting fuses Use smaller lateral fuses Use faster breakers or reclosers Raise the nominal voltage. Reduce faults. Momentary Interruptions Use an instantaneous reclose Use fuse blowing Use single-phase reclosers Install extra downstream devices: fuses or reclosers. Use sequence coordination with downstream devices Reduce faults. Sustained Interruptions Use fuse saving Install extra downstream devices: fuses, reclosers, or sectionalizers. Use faulted-circuit indicators. Reduce faults.

All of the options for power quality improvement discussed in this report are utility-side options. It is often more efficient to provide a solution right at the sensitive customers site. Some of the customer solution options are shown in Table 2-2.

2-3

Basic Approaches

Table 2-2 Customer-Side Options for Improving Power Quality3

Voltage Sags A voltage sag is defined as a decrease in root mean square (RMS) voltage magnitude lasting from 0.5 to 30 cycles. Voltage sags usually are caused by a fault in the utility transmission or distribution system. Power-line faults can be caused by animals being caught in power lines, cars striking utility poles, or lightning striking power lines. Most of the faults in the utility system are cleared by the operation of protective devices in the utility system. Substation breakers are typically set up with a reclosing relay to open momentarily during a fault condition and allow the fault to clear. As shown in Figure 2-1, a customer actually would experience a brief interruption if a fault occurred on the supplying distribution feeder. All other customers on adjacent feeders will experience a voltage sag, the duration of which will depend on the recloser clearing time. The
3

IEEE Std. 1100-1999, Recommended Practice for Power and Grounding Electronic Equipment.

2-4

Basic Approaches

depth of the sag will depend on how close the customer is to the fault location, resistance at the fault, and the available fault current. On the other hand, faults on the transmission system will cause a sag that affects all the customers downstream of that location. Figure 2-2 shows a voltage sag that caused the system voltage to fall to approximately 78% of nominal voltage for 3 or 4 cycles. Based on the above discussion, it can be seen that customers can be affected by sags caused by faults at numerous other locations in the system besides the feeder that is supplying them.

Fault on transmission system cleared by transmission line breakers

Transmission Distribution
Feeder breakers

Fault on fused branch cleared by fuse (may also cause feeder breaker to operate)

End user affected by voltage sags

Fault on main feeder cleared by feeder breaker or recloser

Figure 2-1 Example Distribution System Showing Fault Location and Customer Impact

150 100 50 0 -50 -100 -150 0 25 50 75 100 T i me 125 150 175 200

Figure 2-2 Example Voltage Sag During a Fault

2-5

Basic Approaches

Several power quality studies have been performed that characterize the frequency of voltage sags. Table 2-3 shows the distribution of voltage sag magnitudes and durations from a National Power Laboratory (NPL) study4.
Table 2-3 Cumulative Voltage Sag Table for NPL Data With 5-Minute Filter: Number of Sags per Year Magnitude (% of Nominal) 87 80 70 50 10 1 Cycle 126.4 44.8 23.1 15.9 12.2 6 Cycles 56.8 23.7 17.3 14.1 12.0 10 Cycles 36.4 17.0 14.5 12.9 11.7 Duration 20 Cycles 27.0 13.9 12.8 11.8 11.0 0.5 Sec 23.0 12.2 11.5 10.6 10.2 1 Sec 18.1 10.0 9.7 9.4 9.0 2 Sec 14.5 8.0 7.9 7.8 7.5 10 Sec 5.2 4.3 4.3 4.3 4.2

The effects of voltage sags can be reduced in several ways: Reducing faults (for example, trimming trees, installing tree wire, installing animal guards, installing arresters, and performing circuit patrols) Using faster tripping (that is, smaller fuses, instantaneous tripping, faster transmission relays, and other similar options) Improving the voltage for faults at given locations (for example, raising the nominal voltage, using current-limiting fuses, using larger station transformers, and/or using line reactors)

The voltage during the fault (Vsag in Equation 2-1) at the substation bus can be calculated using the following voltage-divider equation:

Vsag =

Zf V Z f + Zs

Eq. 2-1

The source impedance (Zs), the feeder line impedance (Zf), and the prefault voltage (V) are shown diagrammatically in Figure 2-3:

Dorr, D.S., Hughes, B., Gruzs, T.M., Juewicz, E. and J.L. McClaine, Interpreting Recent Power Quality Surveys to Define the Electrical Environment, IEEE Transactions on Industry Applications, Vol. 33, No. 6, November 1997, pp. 14801487.

2-6

Basic Approaches

Fault

Zs

Zf

Figure 2-3 Voltage Divider Illustration Giving the Bus Voltage for a Downstream Fault

The voltage divider equation shows that as the fault is electrically closer to the bus (smaller Zf ), the sag will be deeper. In addition, as the available fault current decreases (larger Zs), the sag becomes deeper. The source impedance includes the transformer impedance plus the subtransmission source impedance. (Much of the time, subtransmission impedance is small enough to be ignored.) The impedances used in Equation 2-1 depend on the type of fault it is. For a three-phase fault (this will result in the most severe voltage sag), the positive-sequence impedance is used (Zf = Zf1). For a line-to-ground fault (the least severe voltage sag), the impedance is the loop impedance, which is Zf = (2Zf1+Zf0)/3. The impedances of lines and transformers are summarized in Table 2-4 and Table 2-5. For 15-kV systems, a good approximation is 1 ohm for the substation transformer (which would be a 7-kA to 8-kA bus fault current) and 1 ohm per mile of line (for ground faults). For accuracy, complex division should be used because the impedances are complex, but for back-of-the-envelope calculations, just the magnitude is sufficient.
Table 2-4 Typical Line Impedances (Ohms/Mile) PositiveSequence Impedance 0.486 + j0.643 0.294 + j0.650 0.207 + j0.628 Zero-Sequence Impedance 1.134 + j2.078 0.876 + j1.952 0.720 + j1.849 Ground-Fault Impedance 0.702 + j1.122 0.488 + j1.084 0.378 + j1.035

Line 4/0 phase, 1/0 neutral all aluminum conductor (AAC) 350-kcmil phase, 2/0 neutral AAC 500-kcmil phase, 3/0 neutral AAC

2-7

Basic Approaches

Table 2-5 Typical Transformer Impedances (Z1 = Z0 = Zground fault = jZ%*kV2/MVA) Ohms With Bolted-Bus Fault Currents Given in Parenthesis Transformer Base MVA (OA rating) 10 MVA, 8% 25 MVA, 10% 50 MVA, 15% 12.47 kV j1.24 (5.8kA) j0.62 (12kA) j0.47 (15kA) 24.94 kV j4.98 (2.9kA) j2.49 (5.8kA) j1.87 (7.7kA)

Another way to approximate the voltage divider equation (see Equation 2-2) is to use the available short-circuit current at the substation bus and the available short-circuit current at the fault location:

Vsag = 1
where

If Is

Eq. 2-2

If = the available fault current on the feeder at the fault location, and Is = the available fault current at the substation bus. Note that this can be used for any type of fault as long as the appropriate fault values are used in the equation. If the angles are ignored, the equation will be an approximation (which is usually acceptable). Figure 2-4 shows a profile of the substation bus voltage for faults at the given distance along the line for 12.47 kV and 24.94 kV. The higher-voltage system results in more severe sags for faults at a given location. If the higher-voltage system had a larger transformer (which would be common for a higher-voltage system), then the difference would not be as great. The graph also shows that three-phase faults cause more severe sags.

2-8

Basic Approaches

0 .8
P e r u n it lin e -to -g ro u n d vo lta g e

1 2 .5 kV 2 5 kV

0 .6

0 .4

0 .2
L in e -to -g ro u n d fa u lt T h re e -p h a se fa u lt

0 .0 0 1 2 3 4 5 6
D is ta n ce fro m th e s u b sta tio n , m ile s

Figure 2-4 Substation Voltage Profile for Faults at the Given Distance (Single-Phase and Three-Phase Faults Are Shown for Each Voltage) for a 25-MVA, 10% Transformer and 350-kcmil Feeder

The effect of feeder faults on voltage sags at the substation bus can be estimated with Equation 3.
P(Vsag ) = n Vsag Z S 1 Vsag Zf
Eq. 2-3

where P = annual number of sags per year where the voltage sags below Vsag, Vsag = per-unit voltage sag level of interest (0 to 1for example, 0.7), n = number of feeders off the bus,

= feeder mains fault rate per mile per phase, including faults on laterals and including
both temporary and permanent faults, Zf = feeder impedance in ohms/mile (normally use [2Z1+Z0]/3 for ground faults), and Zs = source impedance in ohms. The distribution of voltage sags based on Equation 2-3 is shown in Figure 2-5 for some common parameters. 2-9

Basic Approaches
25 A n n u a l n u m b e r o f sa g s b e lo w th e x-a xis va lu e

20

15
25 kV n=4

10

n=2

5
12 kV

0 0 .0 0 .2 0 .4 0 .6 0 .8 P e r unit vo lta g e m a g nitud e

Figure 2-5 Cumulative Distribution of Substation Bus Voltage Sags Per Year for the Given (25-MVA, 10% Transformer, 350-kcmil Feeder, n = 2 or 4 Feeders off the Bus, =1 Faults/Phase/Mile of Mains/Year, Assumes Line-to-Ground Faults Only)

Several points are noted from this analysis on voltage sags: The most important portion of the circuit (as far as circuit improvement, maintenance, or application of current-limiting fuses is concerned) is the first mile or two from the station (for 15-kV systems). Exposure beyond the first two or three miles does not matter (for bus sags). Sags are more severe on higher-voltage distribution systems (for example, 24.94 kV). Faults farther from the substation will cause the station bus voltage to drop (the substation transformer is a higher impedance relative to the line impedance). For 24.94 kV, exposure as far as 5 miles from the station is significant. Three-phase faults cause more severe sags than single-line-to-ground faults. The number of sags on a bus is directly proportional to the number of feeders off the bus. A lower station transformer impedance (a bigger transformer or lower percent impedance) will improve voltage sags at the station bus. It does not matter whether a substation bus tie is open or closed. If it is open, only half of the feeders will be affected by a fault. Faults that do occur cause more severe sags because the effective source impedance is higher. These effects tend to cancel each other. Raising the nominal voltage will improve the voltage seen by customers during a fault. Say that a fault drops the voltage to 0.8 per unit, and the prefault voltage was 1.0 per unit. If the prefault voltage were 1.1 per unit, the voltage during the sag would be 0.88 per unit. This is not a big difference, but for equipment that is sensitive to sags to 0.7 to 0.85 per unit, it might significantly reduce the number of trip-outs.

2-10

Basic Approaches

Momentary Interruptions Momentary interruptions (also known as momentaries) are primarily the result of utility practice of clearing temporary faults by isolating the faulted circuit through the use of reclosers or reclosing circuit breakers. The devices are usually on the distribution system, but momentaries also can occur for some customers for faults on the subtransmission system. Terms for shortduration interruptions include short interruptions, momentary interruptions, instantaneous interruptions, and transient interruptions, all of which are used with more or less the same meaning. The dividing line for duration between sustained and momentary interruptions is most commonly thought of as 5 minutes (1 minute is also common). Table 2-6 shows the average number of momentaries based on several monitoring studies and surveys. The number of momentaries is highly variable from circuit to circuit and utility to utility. For example, in the Edison Electric Institute (EEI) survey5, the average of the utility averages is 7.9 momentaries, but the average ranged from 1.4 momentaries at the best utility to 19.1 momentaries at the worst. Weather is obviously an important factor, but so are exposure and utility practices.
Table 2-6 Average Number of Momentary Interruptions Found in Several Studies Average Number of Momentary Interruptions per Year 4.6 10.0 3.3
9

Study/Survey EPRI DPQ monitoring study6 NPL monitoring study7 Canadian Electrification Association (CEA) monitoring study8 Institute of Electrical and Electronics Engineers survey Edison Electric Institute (EEI) survey
10

6.6 7.9

1998 EEI Reliability Survey from minutes of the 8th meeting of the EEI Distribution Committee, Mar. 28 31,1999. 6 Dorr, D.S., Hughes, B., Gruzs, T.M., Juewicz, E. and J.L. McClaine, Interpreting Recent Power Quality Surveys to Define the Electrical Environment, IEEE Transactions on Industry Applications, Vol. 33, No. 6, November 1997, pp. 14801487. 7 Dorr, D.S., Hughes, B., Gruzs, T.M., Juewicz, E. and J.L. McClaine, Interpreting Recent Power Quality Surveys to Define the Electrical Environment, IEEE Transactions on Industry Applications, Vol. 33, No. 6, November 1997, pp. 14801487. 8 Dorr, D.S., Hughes, B., Gruzs, T.M., Juewicz, E. and J.L. McClaine, Interpreting Recent Power Quality Surveys to Define the Electrical Environment, IEEE Transactions on Industry Applications, Vol. 33, No. 6, November 1997, pp. 14801487. 9 IEEE Std. 1366-1998, Trial Use Guide for Electric Power Distribution Reliability Indices. 10 1998 EEI Reliability Survey from minutes of the 8th meeting of the EEI Distribution Committee, Mar. 28 31,1999.

2-11

Basic Approaches

Momentaries can be improved in several ways, including: Reducing faults (trimming trees, using tree wire, using animal guards, installing arresters, and performing circuit patrols) Use faster reclosing Limiting the number of customers interrupted (using single-phase reclosers, using extra downstream reclosers, and/or not using fuse saving)

Sustained Interruptions Before the proliferation of sensitive electronic equipment, the major power quality problem was sustained interruptions, which are generally defined as interruptions lasting longer than 1 to 5 minutes. The major cause of these interruptions is faults on the distribution system. The fault causes a fuse, breaker, recloser, or sectionalizer to lock out the faulted section. Many utilities use reliability indices to track the performance of the utility (or region or circuit). Most regulated, investor-owned utilities are required by regulators to report their reliability indices. Nationally, the trend is to move to performance-based rates where performance can be penalized or rewarded based on sustained interruptions as quantified by the reliability indices. Some utilities also use indices as the basis for managerial and/or general bonuses. Most co-ops and municipal utilities do not track these indices, but it is recommended that they at least be aware of them. Some commercial and industrial customers are starting to ask utilities for their reliability indices before locating a facility within the utilitys service territory.
11 According to a 1995 IEEE survey of U.S. electric utilities , the five most commonly used reliability indices are System Average Interruption Duration Index (SAIDI), System Average Interruption Frequency Index (SAIFI), Customer Average Interruption Duration Index (CAIDI), Average Service Availability Index (ASAI), and Momentary Average Interruption Frequency Index (MAIFI). These indices are defined as follows:

SAIDI: This index (see Equation 2-4) gives an indication of the average length of time that a customer is without power.

SAIDI =

r N
i

NT

Eq. 2-4

SAIFI: This index (see Equation 2-5) gives an indication of the number of times that a customer is interrupted.

SAIFI =

N
NT

Eq. 2-5

1998 EEI Reliability Survey from minutes of the 8th meeting of the EEI Distribution Committee, Mar. 28 31,1999.

11

2-12

Basic Approaches

CAIDI: This index (see Equation 2-6) gives an indication of the average time required to restore power to an interrupted customer.
CAIDI =

r N N
i i

SAIDI SAIFI

Eq. 2-6

ASAI: This index (see Equation 2-7) represents the percentage of time that a customer has power over a given period of time (usually one year).
ASAI = N T t y ri N i NT t y
Eq. 2-7

MAIFI: This index (see Equation 2-8) represents the average number of times that a customer is subjected to a momentary outage.

MAIFI =
where

ID N
i

NT

Eq. 2-8

i = interruption event, ri = restoration time for each interruption event, IDi = number of interrupting device operations, Ni = number of interrupted customers for each interruption event, NT = total number of customers served, and ty = total number of hours in a year (8760 hours, 8784 hours for leap year). As an indication of the occurrence of sustained interruptions, Table 2-7 shows the averages of reliability indices reported by utilities in 199512. Most of the reported data excluded major storm or major event interruptions. Major storms are excluded because it is thought that performance during storms (which may significantly alter the indices) does not represent the true performance of the reliability of a distribution system. Storms are defined differently by different utilities. (Some use a percentage of customers interrupted in some way, and others use some form of weather-based classification.) If storms are not excluded, the numbers go up, as shown in 13 Table 2-8 . The repair time (CAIDI) and the average total interruption time (SAIDI) increase the most if storm data are included.

IEEE Std. 1366-1998, Trial Use Guide for Electric Power Distribution Reliability Indices. 1998 EEI Reliability Survey from minutes of the 8th meeting of the EEI Distribution Committee, Mar. 28 31,1999.
13

12

2-13

Basic Approaches

Table 2-7 Reliability Data From a 1995 Survey by the Institute of Electrical and Electronics Engineers Quartiles (Less Than the Value for the Given Percentage of Utilities Responding) 25% Average annual frequency of interruptions (SAIFI) Average annual minutes of interruption (SAIDI) Average minutes per interruption (CAIDI) 0.90 55.7 54.7 50% 1.10 90.0 76.4 75% 1.45 138.3 108.0 1.26 116.9 88.3

Reliability Measure

Average

Table 2-8 Reliability Data With and Without Storms from a 1998 Survey by Edison Electric Institute (EEI) Reliability Measure Average annual frequency of interruptions (SAIFI) Average annual minutes of interruption (SAIDI) Average minutes per interruption (CAIDI) Major Storms Excluded 1.35 99.6 85.2 All Interruptions 1.64 253.8 169.8

Sustained interruptions can be improved in several ways, including:

Reducing faults (trimming trees, using tree wire, using animal guards, using arresters, and performing circuit patrols) Finding and repairing faults faster (installing faulted circuit indicators, using an outagemanagement system, and properly staffing crews) Limiting the number of customers interrupted (using more fuses, reclosers, and sectionalizers Interrupting customers only for permanent faults (using reclosers instead of fuses or sectionalizers and using fuse saving)

Reduction of Faults and Targeted Maintenance Faults are the root cause of sags and interruptions. Obviously, if faults are reduced, the incidence of sags and interruptions will be reduced. Incidence levels can be reduced in several ways, including:

Trimming trees

2-14

Basic Approaches

Using animal guards Using arresters Using tree wire (covered wire) Using aerial or underground cable Identifying and replacing poorly performing hardware Performing line patrols, including thermal scanning

Faults may be either temporary or permanent. A permanent fault is one where permanent damage is done to the system. This includes insulator failures, broken wires, or failed equipment such as transformers and capacitors. Virtually all faults on underground equipment are permanent. Most equipment fails to a short circuit. Permanent faults on distribution circuits usually will cause sustained interruptions for some customers. To clear the fault, a fuse, sectionalizer, recloser, or breaker will have to operate to interrupt the circuit. The most critical location is the three-phase mains, because a fault on the main feeder will cause an interruption to all customers on the circuit. A permanent fault will also cause a voltage sag to customers on the feeder and on adjacent feeders (see Figure 2-6). Permanent faults may also cause momentary interruptions for a customer. A common example is a fault on a fused lateral (tap). Often, a practice called fuse saving is used where an upstream breaker or recloser attempts to open before the fuse blows. After the first attempt to save the fuse, if the fault is still there, the breaker allows the fuse to clear the fault. If a fault is permanent, all customers on the circuit will experience a momentary, and the customers on the fused lateral will experience a sustained interruption. If a circuit is interrupted and then reclosed after a delay, a temporary fault does not involve permanent damage to the system and the system will operate normally. Temporary faults make up 60 to 90% of faults on overhead distribution systems. The causes include lightning strikes, conductors slapping together in the wind, tree branches that fall across conductors and then fall off, animals that cause faults and then fall off, and insulator flashovers caused by pollution. Temporary faults are the main reason that reclosing is used almost universally on distribution breakers and reclosers (on overhead circuits). Temporary faults will cause voltage sags for customers on the circuit with the fault and possibly for customers on adjacent feeders. Temporary faults will cause sustained interruptions if the fault is downstream of a fuse and fuse saving is not used successfully. For faults on the feeder backbone, all customers on the circuit will have a momentary interruption. One way to improve the fault-rate performance is to track the location and type of faults. It is relatively straightforward to track permanent faults. The effort in tracking faults is paid back by targeting maintenance efforts to the most important sections. This is not necessarily the locations with the most faults per mile. The number of customers on a circuit and the type of customers on a circuit are important considerations. For example, a suburban circuit with many hightechnology commercial customers should warrant different treatment than a rural circuit with fewer, mostly residential and agricultural customers. How this is weighted depends on the utilitys philosophy.

2-15

Basic Approaches

Outline of Sag Exposure Momentary Exposure

Customer Location

Sustained Interruption Exposure

Figure 2-6 One-Line Diagram Showing the Boundaries Where Faults Will Cause Sags, Momentaries, and Interruptions for the Customer Shown

Using Outage Databases

Outage databases can help identify problems leading to faults. They can help identify which circuits to target, which areas have the most problems with trees, which areas have the most problems with animals, and so on. Outage databases can also help judge the effectiveness of improvement programs. Without accurate and thorough outage databases, there is no evidence to justify reliability improvement programs. The best strategy is to provide relatively simple database entry codes and remarks fields. Guidelines and training for entering the outage data should be provided. Ensure that the codes are entered accurately by using audits, assessments, and budgetary motivation. Regional budgets for reliability improvement programs are often based on the reliability information extracted from the outage database. Examples of reliability based budget allocations include underground cable replacement, vegetation maintenance, and worst circuit programs. If the correct failure codes are not used in the outage database, then a region may lose money for certain reliability and asset management programs. 2-16

Basic Approaches

All outages contain an element of mystery. The initial coding of outages is based on the best judgment and training of both the first responder and the employee who enters the data. However, the initial coding often does not identify the primary root cause. One of the major goals of an outage follow-up process is to identify the primary root cause. Most outages have more than one root cause. The primary root cause is the root cause for which a utility can implement economical corrections within a reasonable period of time. The purpose of codes and coding conventions is to identify the primary root cause of each outage. Most outage databases provide standard information about the outage such as date, time, customers affected, and other information. For reliability analysis however, the code fields and remarks field provide information that is critical to modern asset management strategies. Having accurate and precise outage code systems increases the usefulness of outage databases. Weather is a common outage cause code, but what does that mean? If a tree knocks down a distribution line during a storm, is that weather? How do you differentiate between lightning and tree-caused outages during a thunderstorm? Another common blunder that crews make is tagging a cause as cutout when the cause was really something downstream of the cutout, and the fuse operated to clear the fault (the cutout operated properly). As much as possible, good outage code systems should separate the root cause of the outage from the weather, the protective device that operated, and the equipment affected. Use a separate category for weather (and indicate major storm separately). Also, try to have codes that reveal deficiencies: these may include inadequate clearances, deteriorated equipment, missing animal protection, or low insulation. Duke Powers system (other utilities use similar systems) provides a good way to characterize an outage by specifying four code fields for an outage: (1) interrupting device, (2) cause or failure mode, (3) equipment code, and (4) weather. These codes, used in combination with each other, are quite accurate in describing what is known about the cause of an outage, or in some cases, what is not known about an outage. In more detail, these codes mean: 1. Interrupting deviceFuse, line breaker (Recloser), station breaker, transformer fuse, etc. 2. Cause or failure modeAnimal, tree, unknown, etc. If the cause is not known, or does not fall into a predetermined cause category, then the cause code becomes the failure mode. Often, the failure mode is the only information available. Examples are burned, broken, malfunctioning, and decayed. Failure modes imply equipment failures, so an equipment code is usually a required entry if using a failure mode. Failure modes represent partial information about the cause, in that the root cause may remain unknown, but certain things about the failure are known. For example, if a station breaker fails to trip, then you may have a specific failure mode called fail to interrupt. For the equipment code (see below) enter the code for station breaker. This system is flexible. The same failure mode can be used with other equipment codes to describe other devices that fail to interrupt. If there is no specific failure mode, then a generic failure mode such as malfunctioning or even broken can be used. 3. Equipment codeFor equipment failures, to specify the equipment that failed. 4. WeatherIt is useful to know if an equipment failure or unknown interrupting device outage occurred during lightning or other stormy weather. Animal outages normally occur in fair weather. By studying combinations of time-of-day, weather, and interruption device activity on a utility, it is often possible to classify unknown outages into likely cause categories. For 2-17

Basic Approaches

example, on the Duke Power system, an unknown outage on a tap fuse on a spring morning in fair weather is highly likely to be an animal outage. There is often discussion concerning if weather is an outage cause or a contributing factor. From a reliability perspective, the electric distribution system is designed to withstand moderate to bad weather conditions. As a primary root cause, weather that exceeds the design parameters of the system can be considered a primary root cause. However, these weather causes are the exception. Weather situations that may be considered primary root causes include: lightning, wind loading exceeding design parameters (tornadoes), ice loading exceeding design parameters (example: three inches of ice), or flooding. Even given that, utilities should avoid declaring stormy weather such as wind or lightning as a cause. There may be evidence (burn marks for example) to show that lightning caused an arrester failure, but if there is not, it is better to code the cause as an equipment failure during a lightning storm. Do not use weather as an excuse to explain away outages. Reliability programs can prevent many outages during stormy weather. Properly funded and executed vegetation management programs can cut down on outages during wind. Properly designed overhead lines and equipment protection can substantially reduce outages during lightning. If the utility is not attempting to make the system more outage resistant during stormy weather, it may be thinking of weather as a cause, rather than a contributing factor. Engineers often want to specify additional codes such as vintage year (for equipment failures), overhead/underground indicators, how the outage was restored, and others. Most of these codes can be incorporated in the four main codes given above. First responders will likely ignore the rest. For example, underground cable failures have equipment codes that are obviously underground types. Experience on the Duke Power system has shown that the four codes mentioned above are about the limit that a first responder can accurately handle, even if other codes are requested. Have a generic comment or remarks section to note specifics that might not be clear from the outage codes. This helps with later analysis and allows for keyword searching revealing patterns. The remarks field contains vast amounts of informative if used liberally. Provide at least 256 characters in the remarks field. Encourage first responders and dispatchers to use the remarks field to describe what happened during the outage. Modern database applications can quickly search and count outages that contain certain keywords or phases within the remarks field. These database applications can also filter out records that contain certain keywords. Here are some examples actually used at Duke Power.

There was no code to differentiate between live or dead tree outages in the code system. However, the crews almost always noted in the remarks field if the tree or limbs involved were dead or rotten. This method of counting dead tree outages versus live tree outages was found to be a highly accurate method to determine if the annual danger-tree survey and removal was effective. There was a hypothesis that 3D or smaller transformer fuses were more susceptible to nuisance fuse outages during lightning than 5D or larger transformer fuses. There were a large number of outage records of fuse outages on transformers during lightning, but the fuse size was not a record anyone kept. However, it was discovered that the crew almost always called in the fuse size when they replaced a fuse. The dispatchers put this information in the remarks field. A contextual database search on the remarks field of transformer outages found more than seven thousand cases per year where the transformer fuse size was supplied.

2-18

Basic Approaches

This sample size was sufficient to allow normalization of 3D and 5D transformer outages against the total population of transformers that would have such fuses. The result of this analysis showed that there was no significant difference in the performance of these two fuses.

There was a concern that squirrels were causing significant damage to overhead conductors, connectors, and equipment by literally biting and gnawing on various items. The remarks field was searched to find how widespread this behavior was. Duke Power determined that squirrels will chew and gnaw almost anything, but bare aluminum conductor is their favorite. Also, while widespread, this phenomenon is not a significant reliability problem.

When designing or modifying outage code systems, balance your desire for as much information as possible with the reality that overloading crews with too many options is counterproductive. Stick with clear choices, and dont overwhelm crews with options. Do not use codes that are too generic to provide useful information. Do not use codes that are too complicated for first responders to understand. In either case you get little useful information. A balance must be maintained between being too generic and too complicated. Here are some additional guidelines for outage coding:

SubcodesWhere appropriate, subcodes provide extra information. As an example, it may make sense to have a code or subcode to denote whether the structure had an animal guard. If that adds too much complexity, have crews enter a generic comment to indicate the structural deficiency (the missing animal guard). Also, for animal outages, crews should only mark it as animal if they find the remains of the animal or other proof. UnknownUsing unknown as a cause is better than guessing. A wrong cause code can result in resources being directed to the wrong problem. Equipment failuresEquipment failures are prime candidates for fine-tuning. For example, a subcode could indicate the equipment construction or vintage or even manufacturer. For cables and especially for splices, knowing the manufacturer and vintage can help determine targeted replacement programs. The mode of failure is also an important consideration: Was the equipment broken? Was the failure from decay?

Consider an example outage code for a tree contact during a storm that caused a downed wire that might be tagged as follows: Interrupting device: Breaker Cause: Vegetation [dead tree]

Equipment: Bare conductor [downed wire] Weather: Wind/Rain This clearly separates the cause (the tree) from the impact on the system (downed wires) and notes the weather conditions during the event. The items in brackets denote subcodes that add useful information; in this case, the subcodes reveal that the tree was dead (points to a better hazard-tree program) and that it caused a downed wire. An arrester failure found during a lightning storm might be classified as: Interrupting device: Transformer fuse Cause: Equipment failure 2-19

Basic Approaches

Equipment: Arrester [catastrophic]

Weather: Lightning storm

In this case, we dont really know the cause of the arrester failure; it was probably lightning, but we dont know for sure. We dont even know for sure that the arrester failed during the storm (the arrester may have failed earlier and was only found when crews arrived at the scene). But because the arrester was obviously failed just downstream of the fuse, and the transformer was still operational, the arrester is very likely to have caused the fault. Consider a squirrel outage across a bushing: Interrupting device: Tap fuse Equipment: Bushing [no animal guard] Cause: Animal Weather: Clear

The use of a subcode for bushing that highlights a common deficiency (no animal guard) helps direct resources to repair the deficiency. Consider also the way crews provide the information. Using mobile computers allows the most direct data entry, but crews are prone to incorrectly using the software (either not using it or entering the data to process the menus as quickly as possible). Training and straightforward user interfaces can help. The advantage of computer data entry is that the form can adapt to the scenario at hand and fill in data from the outage management system (like outage start time). Having a crew call back outage information to dispatchers allows the dispatchers to query the crew to make sure the codes are entered consistently. Paper data entry sheets can be relatively easy for crews to interpret but limit flexibility for sub-codes and options. Utility culture should also encourage accurate outage codes. Provide a document to show how all outage codes are supposed to be used, and provide training. These guidelines and code sheets should be provided to everyone who has anything to do with entering outage codes. At Duke Power, dispatchers enter outage codes that first responders call in. At the 24-hour center, dispatchers are graded on how accurately the code guidelines are followed. Local reliability engineers or technicians are the code police. Each day, these employees review all the outage records in their location for the past twenty-four hours or over the weekend. They make sure that the code guidelines are followed. If there is something missing, unclear, or contradictory about the information in the outage record, the reliability engineer tracks down the information and completes the entry correctly. The code police are also assessed quarterly on the accuracy of outage records in their location. Those who police outage data make sure, to the best of their ability, that all outages in their location are reported accurately and completely. Therefore, these individuals should be rewarded for the accurate and complete reporting of outages. To avoid conflicts of interest, these individuals should not have annual outage indices such as SAIFI or SAIDI on their scorecards.

2-20

3
PROTECTION STRATEGIES
Improvements in the application of overcurrent protective devices can improve utility reliability and power quality by minimizing the effect of faults on customers. Options include using more protective devices that can sectionalize and restore circuits faster (automation), using more protective devices (more fuses, more reclosers), reclosing faster, improving coordination, and so on. An advantage of optimizing protection as a way of improving reliability is that many options are low cost-utilities can improve performance by changing relay or recloser settings or applying low-cost fuses more appropriately.

Single-Phase Protective Devices

Many overcurrent protective devices for distribution systems are single-phase or are available in single-phase versions, including reclosers, fuses, and sectionalizers. Single-phase protective devices reduce the number of customers subjected to momentary and sustained interruptions (Example: If a fault occurs on phase A, single-phase customers on phases B and C do not have an interruption). However, they can have some negative effects when applied on three-phase circuits such as:

Creates possibility of ferroresonance Causes single-phasing of three-phase motors May not clear fault because of backfeed from the other two phases

Because they can greatly reduce the number of customer interruptions experienced on a feeder, single-phase protective devices are should be used in the following manner:

All single-phase taps Three-phase mains and taps if caution in application is used, especially when large, threephase customers are downstream

Reclosers with single-phase trip and three-phase lockout are preferred on three-phase mains serving a mixture of single- and three-phase customers. Industry Usage Single-phase protective devices are used widely on distribution systems. On single-phase laterals, the single-phase devices are almost universally fuses. On long, single-phase laterals, single-phase reclosers are sometimes used. Most utilities also use fuses for three-phase laterals. 3-1

Protection Strategies

The utilities that do not use single-phase devices most often cite the problem of single-phasing three-phase customers motors. Some utilities routinely use single-phase reclosers on three-phase circuits and, in certain circumstances, even in the substation. Estimating the Improvement Because single-phase protective devices on single-phase laterals are used so widely, and the benefits are universally accepted, this practice will not be discussed here. For single-phase devices used on three-phase systems, estimating the effect on individual customers is relatively easy to calculate using the number of phases that are faulted on average, as shown in Table 3-1. Overall, using single-phase protective devices will cut the average number of interruptions in half. This assumes that all customers are single-phase and that the customers are evenly split between phases.
Table 3-1 Effect on Interruptions When Using Single-Phase Protective Devices on Three-Phase Circuits Percent of Customers Affected Using Single-Phase Devices 33 67 100 TOTAL

Fault Type

Percent of Faults 70 20 10

Weighted Effect

Single-phase Two-phase Three-phase

23% 13% 10% 47%

If three-phase customers are downstream of the single-phase devices, they too will generally have an improvement. Three-phase customers will have many single-phase loads, and the loads on the unfaulted phases will not be affected by the fault. Three-phase devices may also ride through an event caused by a single-phase fault (although motors may heat up because of the voltage unbalance as discussed in the following section entitled Single-Phasing of Three-Phase Motors). Three-phase loads are often controlled by single-phase devices (which are often the most sensitive element). The effect on three-phase customers depends on how loads are connected and on the transformer connection, as shown in Table 3-2. In general, the more phases that are affected, the less severe the voltage drop is on those phasesone situation is not always better than the other. An unbalanced voltage may cause three-phase devices to trip. It depends on the type and design of the device and its controls.

3-2

Protection Strategies Table 3-2 Line-to-Ground and Line-to-Line Voltages With One Phase Interrupted Phase-to-Neutral Voltages Per unit VA Wye-wye Delta-wye 0.00 0.58 VB 1.00 1.00 VC 1.00 0.58 VAB 0.58 0.88 Phase-to-Phase Voltages Per unit VBC 1.00 0.88 VCA 0.58 0.33

Transformer Connection

Single-Phasing of Three-Phase Motors One of the main concerns of using single-phase protective devices on three-phase circuits is that motors can overheat and be damaged by single-phasing. Motors are extremely sensitive to voltage unbalance. Motors have relatively low impedance to negative-sequence voltage; therefore, a small negativesequence component of the voltage produces a relatively large negative-sequence current. Consequently, the effect is magnified, so a small negative-sequence voltage results in a significantly larger percentage of unbalanced current than the percentage of unbalanced voltage. The negative-sequence component of stator current is directly related to this unbalance and sets up a counter-rotating flux field in the machine. This, in turn, causes local heating in the rotor iron. The capability of machines to withstand heating caused by unbalanced currents is typically expressed in terms of a constant and is supplied by the manufacturer of the machine. Loss of one or two phases is a large unbalance. For one open phase, the phase-to-phase secondary voltages become 0.57, 1.0, 0.57 per unit for a wye-wye-connected transformer and 0.88, 0.88, 0.33 per unit for a delta-wye-connected transformer. In either case, the negativesequence voltage is 0.66 per unit. With such a high unbalance, the motor can get hot quickly. The negative-sequence impedance of a motor is roughly 15%, so for a 66% negative-sequence voltage, the motor will draw a negative-sequence current of 440%. National Electrical Manufacturers (NEMA) MG-1, Part 20, Section III entitled Large MachinesInduction Motors and Part 21 entitled Large MachinesSynchronous Motors use 14 the same derating curve for operation under unbalanced voltage . The derating curve is shown in Figure 3-1.

14

NEMA MG-1, Standard for Motors and Generators, National Electrical Manufacturers Association, 1998.

3-3

Protection Strategies
Motor Derating for Unbalanced Voltages
1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.0%

Motor Derating Factor (pu)

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.5%

5.0%

Voltage Unbalance (%)

Figure 3-1 Derating Curve for Induction/Synchronous Motors (NEMA MG-1 Part 20 and 21)

In addition to motors, voltage unbalance causes three-phase power electronic loads such as adjustable-speed drives (ASDs) to draw unequal current in each phase. This effect is more pronounced in smaller drives (< 15 hp) where the percent current unbalance may be 20 times the percent voltage unbalance. High phase currents can result in overcurrent trips and excessive dc bus ripple. High ripple can reduce the life of the dc bus capacitor of the ASD. Most service agreements with customers state that it is the customers responsibility to protect their equipment against single-phasing. The best way to protect motors is with a phase-loss relay. Nevertheless, some utilities take measures to reduce the possibility of single-phasing customers motors, and one way to do that is to limit the use of single-phase protective devices. Other utilities are more aggressive in their use of single-phase protective equipment and leave it up to customers to protect their equipment. Ferroresonance Ferroresonance is a special form of resonance that involves the magnetizing reactance of a transformer and the system capacitance. A common form of ferroresonance occurs during singlephasing of three-phase distribution transformers15. Ferroresonance most commonly occurs on cable-fed transformers because of the high capacitance of the cables. The transformer connection is also critical for ferroresonance. An ungrounded primary connection leads to the highest magnitude ferroresonance. During single-phasing (which occurs when a line crew energizes or de-energizes the transformer with single-phase cutouts at the cable riser pole, as shown in Figure 16 3-2 ), a ferroresonant circuit may be set up between the cable capacitance and the transformer reactance. If these impedances are in the range of magnitudes where ferroresonance can occur,
R.H. Hopkinson, Ferroresonant Overvoltage Control Based on TNA Tests on Three-Phase Delta-Wye Transformer Banks, IEEE Transactions on Power Apparatus and Systems, Vol. 86, pp. 12581265, Oct. 1967. 16 An Assessment of Distribution System Power Quality: Volume 3: 2: Library of Distribution System Power Quality Monitoring Case Studies, EPRI, Palo Alto, CA: 1996. TR-106294-V3.
15

3-4

Protection Strategies

then high voltages up to 5 per unit can occur on the open legs of the transformer. The most common problem that can occur is overheating of metal-oxide surge arresters.
150 100 50
% Volts

0 -50 -100 -150 0 10 20

Time (mSeconds)

30

40

50

60

70

Figure 3-2 Example of Ferroresonance Measured During a Manual, Single-Pole Switching Operation as Part of the EPRI Distribution Power Quality Project

Ferroresonance has the following characteristics:

Disappears with load as little as a few percent of the transformer rating. Is more probable at higher primary voltages. Shorter cable lengths are required for ferroresonance. Resonance can even occur with no cable, solely because of the internal capacitance of the transformer. Is more probable in smaller transformers. Is more probable in low-loss transformers.

The probability of ferroresonance occurring is determined by the capacitance (cable length) and by the core losses (see Figure 3-3)17. The core losses are an important part of the ferroresonant circuit. The ferroresonant circuit can be visualized using the simplification shown in Figure 3-4. The transformer-magnetizing branch can be viewed as the core-loss resistance in parallel with a switched inductor. When the transformer is unsaturated, the switched inductance is open, and the only connection between the capacitance and the system is through the core-loss resistance. When the core saturates, the switch closes and the capacitive charge dumps into the system. This condition occurs every cycle. If the core loss (or the resistive load on the transformer) is large enough, the charge on the capacitor is drained off before the next cycle, and ferroresonance does not occur. If the cable susceptance is greater than the transformer core loss conductance, then ferroresonant overvoltages may occur.

An Assessment of Distribution System Power Quality: Volume 3: 2: Library of Distribution System Power Quality Monitoring Case Studies, EPRI, Palo Alto, CA: 1996. TR-106294-V3.

17

3-5

Protection Strategies

Figure 3-3 Ferroresonance on a Cable-Fed Transformer With an Ungrounded High-Side Connection

transformer magnetizing reactance

cable capacitance

Figure 3-4 Equivalent Circuits of Ferroresonance on a Transformer With an Ungrounded High-Side Connection

Table 3-3 shows what type of transformer connection is susceptible to ferroresonance. To avoid ferroresonance on ungrounded, wye-delta-connected transformers, some utilities temporarily ground the wye on the primary side of ungrounded wye-delta connections during switching operations. Ferroresonance can occur on transformers with a grounded primary connection if the windings are on a common core such as the five-legged core transformer (the magnetic coupling between phases completes the ferroresonant circuit18).
Table 3-3 Transformer Primary Connections Susceptible to Ferroresonance Connections Susceptible Not Susceptible

Ungrounded wye Delta Grounded wye with three-, four-, or fivelegged core construction

Grounded wye made of three individual units or units of triplex construction Open wye, open delta

Ferroresonance is a function of the cable capacitance and the transformer no-load losses. The lower the losses relative to the capacitance, the higher the ferroresonant overvoltage can be. For

Smith, D.R., Swanson, S.R and J.D. Borst, Overvoltages With Remotely Switched Cable-Fed Grounded WyeWye Transformers, IEEE Transactions on Power Apparatus and Systems, Vol. 94, pp. 18431853, Sept/Oct. 1975.

18

3-6

Protection Strategies

transformer configurations that are susceptible to ferroresonance, ferroresonance can occur under 19 the condition shown in Equation 3-1 :
BC PNL Eq. 3-1

where BC = percent capacitive susceptance (BC = cable kvar/transformer kVA) and PNL = percent core loss on the transformer nameplate base. If the condition above is exceeded, then ferroresonance might occur. Solutions to ferroresonance include:

Use a higher-loss transformer. Use higher-rated arresters at the transformer. Use a three-phase switching device instead of a single-phase device. Use a transformer connection not susceptible to ferroresonance. Limit the cable length feeding the transformer to meet the criteria given above.

Ferroresonance most commonly happens when switching an unloaded transformer. It also usually happens with manual switching; ferroresonance can occur because a fault clears a singlephase protective device, but it is rare. The main reason that ferroresonance is unlikely for most situations where a single-phase protective device is used is that there is almost always enough load on the circuit to prevent ferroresonance. Also, if several customers are on a section, the transformers will generally have different characteristics, which will usually act to lower the probability of ferroresonance (and ferroresonance is less likely with larger transformers). Because ferroresonance will be uncommon with single-phase protective devices, it is usually not a factor in protective device selection. The main place where caution may be warranted is on small, three-phase transformers that may be switched unloaded (especially at 24.94 kV or 34.5 kV). Backfeed to Faults During a line-to-ground fault where a single-phase device opens, backfeed through a three-phase load can cause backfeed to the fault, as shown in Figure 3-5. It is a common misconception that this type of backfeed can only happen with an ungrounded transformer connectionbackfeed can also occur with a grounded, three-phase connection.

Walling, R.A., Barker, K.D., Compton, T.M and L.E. Zimmerman, Ferroresonant Overvoltages in Grounded Wye-Wye Pad-Mounted Transformers With Low-Loss Silicon-Steel Cores, IEEE Transactions on Power Delivery, Vol. 8, pp. 16471660, July 1993.

19

3-7

Protection Strategies
distribution substation single-phase device open three-phase transformer(s)

IF

RF line-to-ground fault

Figure 3-5 Backfeed to a Downed Conductor

The general equations for the backfeed voltage and current based on the sequence impedances of the load are20:
IF =

( A 3Z 0 Z 2 ) V
3Z 0 Z 1 Z 2 + R F A

Eq. 3-2

VF = RF I F

Eq. 3-3

where A = Z 0 Z1 + Z1 Z 2 + Z 0 Z 2 , Z1 = positive-sequence impedance of the load, Z2 = negative-sequence impedance of the load, Z0 = zero-sequence impedance of the load, RF = fault resistance, IF = backfeed current, VF = backfeed voltage, and V = line-to-neutral voltage. The line and source impedances are left out of the equations because they are small relative to the load impedances. Under an open circuit with no fault (RF = ), which is also the condition found when a single-phase switch is operated, the backfeed voltage becomes:
VF =

( A 3Z 0 Z 2 ) V
A

Eq. 3-4

D.R. Smith, Impact of Distribution Transformer Connections on Feeder Protection Issues, Texas A&M Annual Conference for Protective Relay Engineers, Mar. 2123, 1994.

20

3-8

Protection Strategies

For an ungrounded transformer connection (Z0 = ), the equation for backfeed current becomes:
IF =

(Z1 2Z 2 ) V 3Z 1 Z 2 + R F (Z 1 + Z 2 )

Eq. 3-5

The backfeed will be different depending on the transformer connection and the load:

A grounded-wye-to-grounded-wye-connected transformer will not backfeed the fault with no load or balanced line-to-ground loads (no motors). It will backfeed the fault when line-to-line load is used, especially motors. An ungrounded primary transformer will backfeed the fault under no load. It may not be able to provide much current, but there can be significant voltage on the conductor. Motor load will increase the backfeed current available.

Regardless of the transformer connections, the backfeed current (for a bolted fault on the primary) is principally a function of the motor load as shown in Figure 3-6. The voltage on the open phase for an open circuit on the primary is a function of the portion of load that is motors (see Figure 3-7). The open-circuit voltage is likely to be higher with a delta-wye transformer.

80

A va ila b le B a ckfe e d C u rre n t [A ]

60

40

20

0 0 200 400 600 800 1000 M o to r kV A

Figure 3-6 Available Backfeed Current on a 12.47-kV Circuit (Wye-Wye- or Delta-Wye-Connected Transformer)

3-9

Protection Strategies

Wye-Wye

Delta-Wye

O p e n Circ u it B a ckfe e d V o lta g e [kV ]

O p e n Circu it B a ckfe e d V o lta g e [kV ]

0 0 20 40 60 80 100 P e rce nta g e o f L o a d tha t is M o to rs

0 0 20 40 60 80 100 P e rce nta g e o f L o a d tha t is M o to rs

Figure 3-7 Available Backfeed Voltage on a 12.47-kV Circuit

The following facts should be considered when evaluating the possibility of a backfeed occurring:

The backfeed voltage is sufficient to be a safety hazard to workers or the public (for example, in a wire-down situation). The available backfeed is not a stiff enough source to maintain an arc of significant length. It also will not maintain a high-current arc, which limits damage at the fault location. It may maintain a low-level arcing and sputtering fault.

Based on these conclusions, while backfeeding raises some concerns, it is not an insurmountable obstacle. Utilities regularly fuse three-phase taps and do not experience enough problems to justify other approaches. Single-Phase Reclosers With Three-Phase Lockout Many of the single-phase reclosers and recloser controls now come with a controller option for a single-phase trip and three-phase lockout. For single-phase faults, only the faulted phase opens. For temporary faults, the recloser will successfully clear the fault and close back in, so there will only be a momentary interruption on the faulted phase. If the fault is still present after the final reclose attempt (a permanent fault), the recloser will trip all three phases and will not attempt additional reclose attempts. This recloser feature eliminates problems with ferroresonance and single-phasing of motors. Single-phasing motors and ferroresonance cause heating, and heating usually takes many minutes for damage to occur. Short-duration single-phasing that would occur during a typical reclose cycle would not cause heating damage. If the fault is permanent, all three phases are

3-10

Protection Strategies

tripped and locked out, so there is no long-term single-phasing. A three-phase lockout also reduces the chance of backfeed to a downed wire for a prolonged period. Single-phase reclosers have traditionally been available in smaller sizes than three-phase reclosers, but now single-phase reclosers have sufficiently high continuous and interrupting ratings that they could even be used in almost all feeder applications and even many substation applications (see Table 3-4).
Table 3-4 Highest Current Ratings of Single-Phase Reclosers That Have Three-Phase Lockout Capability or Three-Phase Reclosers With Single-Phase Trip Manufacturer, Model Cooper, D 1 Cooper, VXE15 1 Joslyn Trimod, 100 1 Lexington, EV815 3 Voltage (kV) 2.414 2.414 417.1 15.5 Continuous Rating (A) 560 400 800 400 Interrupting Rating (A) 10,000 8000 12,500 8000 Medium Oil Vacuum Vacuum Vacuum Control Hydraulic Electronic Electronic Electronic

Another consideration with single-phase reclosers versus three-phase devices is that a ground relay is often not available. A ground relay provides extra sensitivity for line-to-ground faults. Not having the ground relay is a tradeoff to using single-phase devices. Even if a ground relay is available on a unit with single-phase tripping, if the ground relay operates, it will trip all three phases (which defeats the purpose of single-phase tripping). Single-Phase Protective Devices Summary The use of single-phase protective devices can reduce the number of interruptions to customers by up to 50% when compared to using only three-phase devices. However, this gain in reliability does not come for free. The use of single-phase devices on three-phase circuits can cause problems for three-phase equipment of customers, particularly three-phase motors and threephase adjustable-speed drives. Some transformer connections can allow the two unfaulted phases to backfeed the faulted phase. Using single-phase protection can also lead to ferroresonance in some cases. Careful engineering can reduce these risks to acceptable levels. The use of new, single-phase reclosers with a common control to allow for single-phase tripping but three-phase lockout can reduce the effects of single-phasing three-phase customers while also eliminating concerns about backfeeding faults and ferroresonance.

3-11

Protection Strategies

Instantaneous Reclose
An instantaneous reclose means having no time delay (or extremely short time delay) on the first reclose attempt on breakers and reclosers. Many residential devices such as digital clocks, VCRs, and microwaves can ride through a half-second interruption but not a 5-second interruption (a typical reclose delay used at many utilities), so a fast reclose will help reduce residential complaints. An analysis of instantaneous reclose useage reveals the following positive and negative impacts: Benefits

Reduces the impact of momentary interruptions on residential customers. Reduces customer complaints. Has low cost and can be quickly implemented.

Negatives

May not allow enough time for the fault to clear. Allows small chance of damage to motors after the reclose.

Industry Usage From a power quality point of view, a faster reclose is better. Some customers may notice nothing more than a quick blink of the lights. Florida Power Corporation has reported that a reclosing time of 18 to 20 cycles nearly eliminates complaints21. Another utility that has successfully used the instantaneous reclose is Long Island 22 Lighting Company (now Keyspan) . According to a survey by the Institute of Electrical and Electronics Engineers (IEEE), a time to first reclose of less than 1 second is the most common practice, although the fast reclose practice tends to decline with increasing voltage (see Table 323 5 ).

Dugan, R.C., Ray, A., Sabin, D.D., Baker, G., Gilker, C. and A. Sundaram, Impact of Fast Tripping of Utility Breakers on Industrial Load Interruptions, IEEE Industry Applications Magazine, Vol. 2, No. 3, pp. 5564, MayJune, 1996. 22 T.A. Short and R.A. Ammon, Instantaneous Trip Relay: Examining Its Role, Transmission & Distribution World, Vol. 49, No. 2, Feb. 1997. 23 Working Group on Distribution Protection, W.M. Strang, Chair, Distribution Line Protection Practices Industry Survey Results, IEEE Power System Relaying Committee Report, IEEE Transactions on Power Delivery, Vol. 10, No. 1, pp. 176186, Jan. 1995.

21

3-12

Protection Strategies Table 3-5 Percent of Distribution Lines Using Specified Intervals Between Reclose Attempts Based on an IEEE Survey Interval Between Reclose Attempts (Seconds) 01 15 >5
52.5% 37.7% 9.8% 56.5% 33.7% 9.8% 47.9% 50.0% 2.1% 43.8% 40.6% 15.6%

Voltage Class 5 kV 15 kV 25 kV 35 kV

Effect on Sensitive Residential Devices Table 3-6 shows voltage tolerance information for several residential devices24. A wide range of voltage sensitivity exists, but few of the digital clocks lose memory for a complete interruption that is less than 0.5 seconds. Data indicates that an instantaneous reclose is effective for helping residential customers ride through momentary interruptions without losing many devices. Given the wide variation, some customers will be sensitive to even a 0.5-second interruption. This data is supported by another study where microwave ovens, telephones, and digital clocks were upset 25 by interruptions lasting between 10 and 100 cycles . Note that the instantaneous reclose helps with digital-clock devices whether they be on radio-alarm clocks, VCRs, or microwaves. It does not help at all with computers or other computer-based equipment. While instantaneous reclose is effective for some residential customers, it is not a complete solution for others. Because commercial and industrial customers are heavy users of computers and computer-based equipment, instantaneous reclose is not an effective solution for this customer segment.
Table 3-6 Percent of Devices That Were Able to Successfully Ride Through a Momentary Interruption of Specified Durations Percent of Devices With Successful Ride-Through (%) Device Interruption Duration = 0.5 Seconds
70 60 50 0

Interruption Duration = 2 Seconds

60 0 37.5 0

Interruption Duration = 16.7 Seconds

0 0 0 0

Digital clock Microwave oven VCR Computer

K.B. Bowes, Effects of Power Line Disturbances on Electronic Products, Power Quality, Premier V, pp. 296 310, 1990. 25 E. Le Courtois and D. Deslauriers, Voltage Variations Susceptibility of Electronic Residential Equipment, in proceedings of PQA97 North America, Columbus, Ohio, 1997.

24

3-13

Protection Strategies

Delay Necessary to Avoid Retriggering Faults Sometimes a delayed reclose is necessary if there is not enough time to clear the fault. If a fault arc does not have time to cool, the reclose could retrigger the arc. Whether the arc strikes again is a function of voltage and structure spacings. A 34.5-kV user, Vepco, added a delay to the first reclose because the probability of success of the first reclose was much less than normal for distribution circuits26 . Vepcos success rate for the first reclose after an instantaneous reclose was 25%, which is much less than the 70 to 80% experienced by most utilities. Another item that added to the low success rate of Vepcos 34.5-kV system is that they used a lot of armless design, and the combination of higher voltage and tighter spacings requires a longer time delay for the arc to clear. A verbal report by Gene Baker of Florida Power Corporation to an IEEE working group in 1996 stated that at 12 kV, an arc needs approximately 11 cycles of dead time to clear (so a 0.5-second dead time is sufficient). Equation 3-6 has been suggested as a tool to find the minimum deionization time of an arc based on the line-to-line voltage27:

t = 10.5 +
where

kV 34.5

Eq. 3-6

t = minimum de-ionization time in cycles and kV = rated line-to-line voltage in kV. As can be seen from Equation 3-6, the de-ionization time does not increase dramatically with increasing voltage. Even for a 34.5-kV system, the de-ionization time is 11.5 cycles. Equation 1 is a simplification (separation distances are not included), but it does show that the de-ionization time for arcs is fairly quick. A fast reclose is successfully used on many transmission lines. The reclose time for distribution breakers and reclosers varies by design. A typical time is 0.4 to 0.6 seconds for an instantaneous reclose (meaning no intentional delay). The fastest devices may reclose in as little as 11 cycles. This may prove to be too fast for some applications, so a small delay of 0.1 to 0.4 seconds could be added (especially at 25 kV or 35 kV). Another consideration is that on distribution circuits, other things may affect the time to clear a fault besides the de-ionization of air. If a tree limb or animal causes a temporary fault, time may be needed for the limb or animal to fall off the conductors or insulators. Because of this, it is recommended that if an instantaneous reclose is used, the number of reclose attempts before lockout be at least two. For example, a 0-15-30-second cycle could be used (three reclose attempts): The first reclose attempt is done with no intentional delay, the second is done after a
Johnston, L., Tweed, N.B., Ward, D.J. and J.J. Burke, An Analysis of Vepcos 34.5 kV Distribution Feeder Faults as Related to Through Fault Failures of Substation Transformers, IEEE Transactions on Power Apparatus and Systems, Vol. PAS-97, No. 5, pp. 18761884. 27 Applied Protective Relaying, Westinghouse Electric Corporation, 1976.
26

3-14

Protection Strategies

15-second delay, and the third is done after a 30-second delay. If the fault is still present after the third attempt, the distribution circuit is locked out. If two reclose attempts are used, a 0-30- or 0-45-second cycle is recommended. Reclose Impacts on Motors Large industrial customers with large motors may be worried about a fast reclose. The major concern is damage to motors and their driven equipment. The major problem with reclosing is that the motor voltage will not drop instantly to zero when the utility breaker (or recloser) is opened. The motor will have residual voltage (that is, the voltage magnitude and frequency decay with time). When the utility recloses, the utility voltage can be out of phase with the motors residual voltage. This can cause severe stresses on the motor and its driven load. The decay time is a function of the size of the motor and the inertia of the motor and its load. Motors in the 200- to 2000-hp range typically have open-circuit time constants of 0.5 to 28 2 seconds . The time constant is the time it takes for the residual voltage to decay to 36.8% of its initial value. Reclose impacts are worse with:

Larger motors Capacitor banks (excitation from the capacitor banks can greatly increase the motor decay time)

Synchronous motors and generators have much larger time constants and are more likely to be damaged than induction machines, so more consideration should be taken if large synchronous machines are present. On the vast majority of distribution circuits, reclosing impacts will not be a concern because: Most utility feeders do not have motor loads larger than 500 hp.

Even with feeders with large industrial customers, the non-motor load will be large enough to pull the voltage down to a safe level within the time it takes to do a normal instantaneous reclose (0.4 to 0.6 seconds).

Because of this, it is safe to implement an instantaneous reclose on almost all distribution circuits. One exception would be a feeder with an industrial customer that was 90% of the feeder load, and the industrial customer has several large induction or (especially) synchronous motors. Another exception is a feeder where a large, rotating distributed generator is being applied. In both of these cases, an extra time delay may be needed on the first reclose.

28

G.W. Bottrell, Hazards and Benefits of Utility Reclosing, in proceedings of Power Quality '93, pp. 260275, 1993.

3-15

Protection Strategies

Summary Instantaneous reclosereclosing with no time delay or an extremely short time delaycan reduce power quality complaints from distribution customers. This is particularly true for residential customers whose loads are often able to successfully ride through the short (approximately 0.5-second) outage resulting from an instantaneous reclose. Use of instantaneous reclose on circuits with a high percentage of industrial customers with large induction or synchronous machines should be avoided if possible. Reclosing before the motor voltage has had a chance to decay can cause excessive stress on the motor. Care must also be exercised when applying instantaneous reclose to circuits with a large, rotating distributed generator. These generators need time to be isolated from the power system before a reclose.

3-16

Protection Strategies

Fusing
Fusesthe more we have, the more we isolate faults to smaller chunks of circuitry, and the fewer customers we interrupt. Taps are almost universally fused, primarily for reliability. Fuses make cheap fault finders. We want to have a high percentage of a circuits exposure on fused taps, so when permanent faults occur on those sections, only a small number of customers are interrupted. A number of overlooked scenarios can be improved with better application of fuses:

Unfused taps on the mainline Transformers without a local fuse Arresters upstream of fuses

Many utilities have more taps that are unfused than they realize. If the unfused taps use a smaller conductor size, faults are more likely to burn down such conductors because a circuit breaker will take longer to clear a fault than would a fuse. Because such taps may be on side streets, during patrols, crews may forget to inspect them for damage, increasing the interruption time. So, one way to improve reliability is to make sure that all taps are fused. Consider a common example where an unfused tap crosses the street to feed a riser pole. To protect the small, unfused tap, move the fuses from the riser pole to the pole where the circuit taps off of the mainline. Having a local external fuse to protect a transformer helps improve reliability. If the transformer is a completely self-protected transformer (CSP, see Figure 3-8) with an internal fuse, then an animal across a bushing or other bushing failure will force the tap fuse or upstream circuit breaker or recloser to operate, leaving many more customers interrupted, with much more area for crews to patrol. CSPs on the mainline are especially problematic for reliability indices. A local fuse also helps crews more accurately find and identify the source of the problem. The most common fusing equipment for this application is an expulsion fuse in a cutout, but current limiting fuses are also an option. Because CSPs have lower structural withstand capabilities than conventional transformers, current limiting fuses are appropriate in many locations to provide 29 protection against violent transformer failure (see Short for more information).

29

Short, T.A., Electric Power Distribution Handbook. 2004, Boca Raton, FL: CRC Press.

3-17

Protection Strategies

Figure 3-8 Unfused CSP on a Mainline

Arresters should be placed downstream of fuses if possible. Then, an arrester failure or an animal across the arrester will blow the fuse rather than forcing a mainline protective device operation (interrupting many more customers than if a fuse had operated). If the arrester isolator fails to operate (which can happen), the failure may be extremely hard for the crews to find. A fuse helps localize the failure. Arresters upstream of fuses can cause further problems. If the arrester fails and the isolator operates, crews may reclose the circuit successfully if they do not find the failed arrester. This leaves the equipment unprotected. Worse yet, the failed arrester body may start to track across the bracket. Eventually the arrester bracket will flash over causing a hard-to-find permanent fault. With arresters upstream of a fuse, arrester lead lengths will be longer, and the equipment is not protected as well against lightning surges. This can lead to more lightning-caused equipment failures.

Fuse Saving Versus Fuse Blowing

Fuse Saving A fuse-saving protection scheme uses the instantaneous relay element on a breaker or the fast curve on a recloser to operate on faults before downstream lateral fuses operate. This practice is done on the first shot because after the first shot, the relays or recloser is coordinated to allow the fuse to blow before the breaker or recloser trips a second time. Because most faults on distribution circuits are temporary, fuse saving is used to improving reliability. The use of fuse saving has been the traditional way of operating most distribution circuits.

3-18

Protection Strategies

Benefits

Improves reliability for customers. Clears faults faster:

Reduces the duration of fault-induced voltage sags and thus causes fewer problems for customers on adjacent circuits. Creates fewer problems with conductor burndowns and other problems when experiencing longer-duration faults
Eliminates the need to send crews out to replace fuses for temporary faults.

Negatives

Results in more momentary interruptions. Hurts customers on the mainsespecially commercial and industrial customers. May not be possible to save the fuse if the breaker or recloser is not fast enough.

Fuse Blowing A fuse-blowing protection scheme allows the lateral fuse to blow. This has the advantage of greatly reducing the number of momentary interruptions. The fuse-blowing scheme is also called trip-saving or breaker-saving.
Benefits

Fuse blowing improves power quality for industrial and commercial customers on the circuits.
Negatives

Worsens reliability numbers for sustained interruptionsfor example, System Average Interruption Duration Index (SAIDI) and System Average Interruption Frequency Index (SAIFI). Slows clearing of faults:

Increases duration of fault-induced voltage sags and thus causes more problems for customers on adjacent circuits. Creates more problems with conductor burndowns and other problems when experiencing long-duration faults.
Increases number of crew trips to replace blown fuses.

3-19

Protection Strategies

Economics

There is relatively little expense to change from one scheme to the other. Fuse blowing increases outage management costs because of the increased number of trips required to replace blown fuses.

Bottom Line Neither a fuse-saving nor a fuse-blowing protection scheme is the best choice for all applications. One is better in some applications than others. The best choice depends on many factors, including fusing practices, wire type used, the mix and location of customers on a circuit, and the utility philosophy. It is helpful to review the choice made (even on a circuit-by-circuit basis) because many situations would be better served by a different choice. In addition, several variations also exist that may be considered. Introduction Fuse saving is a protection scheme where a breaker or recloser is used to operate before a lateral fuse. This protection scheme is used because a fuse does not have reclosing capability, but the breaker (or recloser) does. Fuse saving is usually implemented with an instantaneous relay element (or the fast curve on a recloser). The instantaneous element is disabled after the first fault, so after the breaker recloses, if the fault is still there, the system is coordinated so the fuse blows. Because most faults are temporary, fuse saving will prevent a number of lateral fuse operations. The main disadvantage of fuse saving is that all customers on the circuit see a momentary interruption for lateral faults. Because of this, many utilities are switching to a fuse-blowing scheme. The instantaneous relay element is disabled, and the fuse is always allowed to blow. Fuse saving is primarily directed at reducing sustained interruptions; fuse blowing is primarily aimed at reducing the number of momentary interruptions. Refer to the example circuit in Figure 3-9 and then refer to Table 3-7 for a comparison of the sequence of events for each mode of operation (that is, fuse saving and fuse blowing).

Breaker or recloser Lateral Fuse

Fault

Figure 3-9 Sample Circuit for Use in Comparing Fuse Saving and Fuse Blowing During a Fault on a Lateral

3-20

Protection Strategies Table 3-7 A Comparison of the Sequence of Events for Fuse Saving and Fuse Blowing Type of Fault Temporary Fault Sequence of Events for Fuse Saving 1. The breaker operates on the instantaneous curve before the fuse operates. 2. The breaker recloses. 3. The fault is gone, so no other action is necessary. Permanent Fault 1. The breaker operates on instantaneous curve before the fuse operates. 2. The breaker recloses. 3. The fault is still there. 4. The instantaneous operation is disabled, so the fuse operates. 5. Crews must be sent out to fix the fault and replace the fuse. 1. The fuse operates. 2. Crews must be sent out to fix the fault and replace the fuse. Sequence of Events for Fuse Blowing 1. The fuse operates. 2. Crews must be sent out to replace the fuse.

Industry Usage Until the late 1980s, fuse saving was almost universally used. As power quality concerns grew, some utilities switched to a fuse blowing protection scheme. A survey that is done periodically by the Institute of Electrical and Electronics Engineers (IEEE) on distribution protection 30 practices has shown a decrease in the use of fuse saving, as shown in Table 3-8 .
Table 3-8 Results of Surveys by the Institute of Electrical and Electronics Engineers on the Percentage of Utilities That Use Fuse Saving Survey Year 1988 1994 Percent of Utilities 91 71

Working Group on Distribution Protection, W. M. Strang, Chair, Distribution Line Protection Practices Industry Survey Results IEEE Power System Relaying Committee Report, IEEE Transactions on Power Delivery, Vol. 10, No. 1, pp. 176186, Jan. 1995.

30

3-21

Protection Strategies

Another survey performed in 1996 showed a mixture of practices at utilities 31 (see Table 3-9) . A few utilities used fuse blowing because they indicated that fuse saving was not successful. Many of the mixed practices utilities decided their choice of fuse saving or fuse blowing for a specific application on a case-by-case basis. Many of these normally used fuse saving but switched to fuse blowing if too many power quality complaints were received.
Table 3-9 1996 PTI Survey on Line-Protection Methods Line-Protection Method Fuse saving Mixture of fuse saving and fuse blowing Fuse blowing Percent of Utilities 40 33 27

Effects on Momentaries and Sustained Interruptions The change in the number of momentaries (as a result of switching from fuse saving to fuse blowing) can be estimated most simply by using the ratio of the length of the mains to the total length of the circuit, including all laterals. For example, if a circuit has 5 miles of mains and 10 miles of laterals off of that, the number of momentaries after switching to fuse blowing would be 1/3 of the number of momentaries with fuse saving (5/(5+10) = 2/3). This calculation assumes that the mains and laterals have the same fault rate. If the fault rate on laterals is higher (which it often is because of less tree trimming and other factors), the number of momentaries would be even less. From this calculation, it is easy to see how dramatically using fuse blowing can reduce momentaries. No other methods exist that can so easily eliminate 30 to70% of momentaries. The effect on reliability of going to a fuse-blowing scheme is a bit more difficult to estimate. 32,33,34 . This increase in fuse Fuse saving will increase the number of fuse operations by 40 to 500% operations will increase the average frequency of sustained interruptions by 10 to 60%. Note that many variables can change the ratios. One example is given in Figure 3-10.

T.A. Short, Fuse Saving and Its Effect on Power Quality, presented at the EEI Distribution Committee, Phoenix, AZ, March 2831, 1999. 32 T.A. Short, Fuse Saving and Its Effect on Power Quality, presented at the EEI Distribution Committee, Phoenix, AZ, March 2831, 1999. 33 Dugan, R.C. Ray, L.A., Sabin, D.D., Baker, G. Gilker, C. and A. Sundaram, Impact of Fast Tripping of Utility Breakers on Industrial Load Interruptions, IEEE Industry Applications Magazine, Vol. 2, No. 3, pp. 5564, MayJune, 1996. 34 T.A. Short and R.A. Ammon, Instantaneous Trip Relay: Examining Its Role, Transmission & Distribution World, Vol. 49, No. 2, Feb. 1997.

31

3-22

Protection Strategies
140% 120%
100% 100%

127%

Fuse Saving

Fuse Blowing

100% 80% 60% 40% 20% 0%

27%

Sustained Interruptions

Momentary Interruptions

Mains: 10 Miles, Fault Rate=0.5/Mile, 75% Temporary Laterals: 10 Miles Total, 20 Laterals, Fault Rate=2/Mile, 75% Temporary Figure 3-10 Comparison of Fuse Saving and Fuse Blowing on Hypothetical Circuit

Note that the effect on sustained interruptions is not equally applied. Customers on the mains will see no difference in the number of permanent interruptions. Customers on long laterals may have many more sustained interruptions with a fuse-blowing scheme. Coordination Limits of Fuse Saving One of the main reasons that utilities have decided not to use fuse saving is that it is difficult to make it work. Fuses are fast relative to breakers, so where fault currents are high, the fuse can blow before the breaker trips. However, the breaker will likely have received a trip signal from the relay and will open even if the fuse blows. This will result in a fuse operation and a momentary for all customers on the circuit. The most common lateral fuses used are K links, which are relatively fast fuses. Most distribution circuit breakers are 5-cycle breakers. For fuse saving to work, the breaker must open before the fuse blows, so the fuse needs to survive for the time it takes the instantaneous relay to operate (approximately 1 cycle maximum) plus the 5 cycles for the breaker. As an illustration, Figure 3-11 shows the limit of coordination of a 5-cycle breaker and a 100-K fuse. Fuse saving will only coordinate for fault levels less than 1360 A. Smaller fuses will be worse. Note that the breaker time is coordinated with a damage curve of the fuse. The damage curve is 75% of the operating time of the minimum melt curve. Not all protection engineers coordinate using the damage curve, but it provides some conservatism.

3-23

Protection Strategies
1 100 K Fuse

Time [sec]

0.1

5-cycle breaker + 1 cycle relay time

Fuse Damage Curve

0.01 100

1000 Current [A]

10000

Figure 3-11 Coordination of a 100-K Lateral Fuse With a 5-Cycle Breaker

3-24

Protection Strategies

Table 3-10 and Table 3-11 show the limits of coordination of several common lateral fuses for a standard breaker (5 cycle) and a fast breaker/relay combination (3-cycle breaker and -cycle relay). Table 3-12 and Table 3-13 translate these fault currents into distances from the substation at 12.47 kV (assuming an 8-kA fault level at the substation). Note that only the larger fuses shown (>100 A) will coordinate for significant portions of the feeder. Smaller fuses used as second- and third-level fuses would not coordinate over the length of most feeders. The situation is even worse at higher voltages. At 24.94 kV, the distances in Table 3-12 and Table 3-13 are doubled, so fuse saving is more difficult to achieve for longer distances at higher voltages.
Table 3-10 Maximum Fault Currents for Fuse Saving Coordination for Several Common Fuse Links for a 5-Cycle Breaker and a 1-Cycle Relay Time K links Critical Current Size (Amps) 20 K 25 K 30 K 40 K 50 K 65 K 80 K 100 K 140 K 200 K 254 323 398 520 665 816 1078 1354 2162 3401 20 T 25 T 30 T 40 T 50 T 65 T 80 T 100 T 140 T 200 T Size (Amps) 433 552 699 896 1125 1428 1790 2277 3447 5436 T links Critical Current

3-25

Protection Strategies Table 3-11 Maximum Fault Currents for Fuse Saving Coordination for Several Common Fuse Links for a 3-Cycle Breaker and a -Cycle Relay Time K links Critical Current Size (Amps) 20 K 25 K 30 K 40 K 50 K 65 K 80 K 100 K 140 K 200 K 332 424 522 682 875 1070 1407 1763 2823 4409 20 T 25 T 30 T 40 T 50 T 65 T 80 T 100 T 140 T 200 T Size (Amps) 565 723 920 1175 1479 1878 2346 2975 4522 7122 T links Critical Current

Table 3-12 Minimum Distance From the Substation Where the Given Fuse Coordinates for Fuse Saving for a 5-Cycle Breaker and a 1-Cycle Relay Time (12.47 kV, 500 kcmil Phase, 4/0 Neutral) (For 24.94 kV, Multiply Distances by 2) K Links Critical Distance Size (Miles) 20 K 25 K 30 K 40 K 50 K 65 K 80 K 100 K 140 K 200 K 26.5 20.8 16.9 12.9 10.1 8.3 6.3 5.0 3.2 2.1 20 T 25 T 30 T 40 T 50 T 65 T 80 T 100 T 140 T 200 T Size (Miles) 15.5 12.2 9.6 7.5 6.0 4.8 3.8 3.1 2.1 1.5 T Links Critical Distance

3-26

Protection Strategies Table 3-13 Minimum Distance From the Substation Where the Given Fuse Coordinates for Fuse Saving for a 3-Cycle Breaker and a -Cycle Relay Time (12.47 kV, 500 kcmil Phase, 4/0 Neutral) (For 24.94 kV, Multiply Distances by 2) K Links Critical Distance Size (Miles) 20 K 25 K 30 K 40 K 50 K 65 K 80 K 100 K 140 K 200 K 20.3 15.9 12.9 9.9 7.7 6.3 4.8 3.9 2.5 1.7 20 T 25 T 30 T 40 T 50 T 65 T 80 T 100 T 140 T 200 T Size (Miles) 11.9 9.3 7.4 5.8 4.6 3.7 3.0 2.4 1.7 1.3 T Links Critical Distance

If smaller K linkssuch as 100-K and 65-K fuses (the most common lateral fuses used)are used, then fuse saving is not going to work very well. A strong argument exists for going to a fuse-blowing scheme. No sense can be found in having a momentary every time a fuse blows (which is what happens when the breaker is not fast enough to save the fuse). Long-Duration Faults and Wire Burndowns With Fuse Blowing A drawback of fuse blowing is that for faults on the mains, faults can take a long time to clear. With fuse saving, main-line faults are normally cleared in 5 to 7 cycles on the first shot with the instantaneous element. With fuse blowing, this same fault may last for 0.5 to 1 second. Much more damage at the fault location can occur during this extra time. Some of the problems that have been identified are:

Wire burndowns occur: The heat from the fault-current arc burns the conductor enough to cause it to break and fall to the ground. Damage to inline equipment occurs: The most common problem has been with inline hot-line clamps. If the connection is not good, the fault arc across the contact can burn the connection apart and cause it to break. Extra fault duty placed on substation transformers. Ground faults are more likely to become two- or three-phase faults.

3-27

Protection Strategies

Faults on under-built distribution are more likely to cause faults on the transmission circuit above.

Of all the items on this bulleted list, wire-burndown problems are the most important consideration. Wire burndowns are usually a problem with the use of the following:

Covered wire: Covered wire (also called tree wire or weatherproof wire) will hold an arc stationary. Because the arc is stationary, burndowns will happen faster than with bare wire. Small, bare wire for the mains: Small, bare wire (less than 2/0) is also susceptible to wire burndowns. This situation might exist if laterals are not fused.

Covered wire is widely used to limit tree faults. Several utilities have had burndowns of covered wire circuits when the instantaneous trip was not used or was improperly applied35,36. If a burndown on the main line occurs, it will be a long interruption for all of the customers on the circuit. In addition, it is a safety hazard. After the wire breaks and falls to the ground, the substation breaker may reclose. After the reclosure, the wire on the ground will probably not draw enough fault current to trip the station breaker again. This is a high-impedance fault that is extremely difficult to detect. Covered wire is susceptible to burndowns because when a fault-current arc develops, the covering prevents the arc from moving. The heat from the arc is what causes the damage. Although ionized air is a fairly good conductor, it is not as good as the wire itself, so the arc gets extremely hot. On bare wire, the arc is free to move, and the magnetic forces from the fault cause the arc to move (in the direction away from the substationa phenomenon called motoring). The covering constricts the arc to one location, so the heating and melting is concentrated on one part of the wire. If the covering is stripped at the insulators and a fault arcs across an insulator, the arc will motor until it reaches the covering, and the burndown can occur there. Lightning, a tree branch, or an animal can be the original cause of the arc. Burndowns are most associated with lightning-caused faults. (Its the fault-current arc, not the lightning, that causes the damage.) Using a fuse-blowing scheme can increase burndowns because the fault duration is much longer on the time-delay relay elements than on the instantaneous element. The wire damage is a function of the duration of the fault and the current magnitude. Burndown damage characteristics are much lower than annealing curves that are published for conductors. As an illustration of the damage and the protection provided by the time-delay relays, Figure 3-12 shows the threshold of 37 damage for a 336.4-kcmil, covered, all-aluminum conductor (data from ). The relay curve shown is a ground relay using a CO-11 (extremely inverse) with a time dial = 5 and a pickup = 300 A. An instantaneous relay characteristic is also shown. The instantaneous element will protect the wire up to approximately 4000 A. If the instantaneous element is removed (to use a fuse-blowing scheme), the ground relay will not protect the wire from damage for any current
T.A. Short and R.A. Ammon, Instantaneous Trip Relay: Examining Its Role, Transmission & Distribution World, Vol. 49, No. 2, Feb. 1997. 36 P.P. Barker and T.A. Short, Findings of Recent Experiments Involving Natural and Triggered Lightning, panel session paper presented at 1996 Transmission and Distribution Conference & Exposition, Los Angeles, CA, September 1620, 1996. 37 Lee, R.E., Fritz, D.E., Stiller, P.H., Kilar, L.A. and D.F. Shankle, Prevention of Covered Conductor Burndown on Distribution Circuits, American Power Conference, 1980.
35

3-28

Protection Strategies

level. The burndown-damage curve shown is the threshold of damagean actual wire breakage will be higher than this by some amount.
100

10

Ground Relay CO-11, TD=5.0 Pickup = 300 A

Time [sec]

Threshold of damage for covered 336.4-kcmil all-aluminum conductor

0.1

5-cycle breaker + 1 cycle relay time

0.01 100

1000 Current [A]

10000

Figure 3-12 Burndown Threshold of Damage Compared to Typical Relay Characteristics

3-29

Protection Strategies

If covered wire is used, the solutions include:

Use fuse saving. Use arc-protective devices (APDs): APDs are sacrificial masses of metal attached to the ends 38 of the conductor where the covering is stripped . Use a delayed instantaneous operation for fuse blowing rather than removing the instantaneous operation. Use a high-set instantaneous curve for fuse blowing rather than removing the instantaneous: Instead of eliminating the instantaneous, set it to trip for faults within the first mile or so of the substation to trip quickly for high-magnitude faults (which are more likely to cause burndowns).

Although not as susceptible as covered wires, bare wires can also have burndowns. Figure 3-13 shows burndown damage information for all-aluminum alloy conductor (AAAC) bare wires along with a 100-K lateral fuse element and the same ground relay element used above39. The fuse protects the wires shown, but the ground relay does not provide adequate protection for wires smaller than #2 AAAC. All-aluminum conductor (AAC) is expected to be slightly worse because the AAAC has a higher breaking strength. The solutions for small bare wire are:

Fuse lateral taps. Use fuse saving. Use a delayed instantaneous for fuse blowing. Use a high-set instantaneous.

Lee, R.E., Fritz, D.E., Stiller, P.H., Kilar, L.A. and D.F. Shankle, Prevention of Covered Conductor Burndown on Distribution Circuits, American Power Conference, 1980. 39 J.A. Lasseter, Burndown Tests on Bare Conductor, Electric Light and Power, pp. 94100, Dec. 15, 1956.

38

3-30

Protection Strategies
100 3/0 1/0 #2 #4 10 Ground Relay CO-11, TD=5.0 Pickup = 300 A

Solid Lines: Average Burndown Dashed Lines: Threshold of Damage Time [sec] 1 3/0 1/0 #2 #4 100 K Maximum Clear 0.1

0.01 100

1000 Current [A]

10000

Figure 3-13 Bare-Wire Burndown Curves for All-Aluminum Alloy Conductor Along With a 100-K Lateral Fuse and a Ground Relay Element

Long-Duration Voltage Sags With Fuse Blowing One of the disadvantages of a fuse-blowing scheme is that voltage-sag durations are longer for temporary faults. This observation is especially true for faults on the three-phase mains that have to be cleared by a phase or ground time-delay element. An example is shown in Figure 3-14, where voltage sag magnitudes and durations are shown for faults at 0, 0.5, 1, 1.5, miles from the substation using a fuse-blowing scheme. For the same circuit with a fuse-saving scheme, all of the faults would have cleared in 0.1 seconds. For a fault at the substation, the duration is tripled. For a fault that is 1 mile from the substation, the duration is quadrupled. The situation is worse for phase-to-phase faults and three-phase faults, because they must be cleared by the phase relays, which are generally set higher.

3-31

Protection Strategies
1 0.9

Voltage Magnitude [per unit]

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.1 1 10

1-mile from the sub

bus fault

Sag Duration [seconds]

The faults are cleared by a ground relay (CO-11, TD = 5, Pickup=300A). For the same circuit with a fuse-saving scheme, all of the faults would have cleared in 0.1 seconds. Figure 3-14 Magnitudes and Durations of Substation Bus Voltage Sags for Ground Faults Applied Every Half Mile on a 12.47-kV Circuit With a Fuse-Blowing Scheme

Optimal Implementation of Fuse Saving To get a fuse-saving scheme to work, it is necessary to get the substation protective device to open before fuses blow. This can be achieved in several ways:

Slow down the fuse: Use big, slow fuses (for example, 140-T or 200-T) near the substation to ensure proper coordination. Use faster breakers or reclosers: If 3-cycle breakers are used instead of the normal 5-cycle breakers, fuse-saving coordination is more likely. Reclosers are even slightly faster than a 3cycle breaker. Limit fault current:

Open station bus ties: An open bus tie will reduce the fault current on each feeder and make fuse saving easier. This practice is the normal operating mode for most utilities. Use a transformer neutral reactor: A neutral reactor will reduce the fault current for single-phase faults.

3-32

Protection Strategies

Use line reactors: This practice will reduce the fault current for all types of faults. The use of line reactors, however, has been an uncommon practice. It has the added advantage of reducing the impact of voltage sags for faults on adjacent feeders. Specify higher-impedance transformers.
Other strategies can also be employed to limit the impact of momentary interruptions:

Use more downstream reclosers: Extra downstream devices will reduce the number of momentaries for customers near the substation. It is important to coordinate reclosers with the upstream device. Use single-phase reclosers Use instantaneous reclose Switch to a fuse-blowing scheme on poorly performing feeders: For a feeder with many momentaries, the instantaneous relay can be disabled for a time period. Poorly performing sections of the circuit can be identified during this time. Once the poorly performing sections are identified and improved, the circuit can be switched back to a fuse-saving scheme.

Optimal Implementation of Fuse Blowing Several strategies can optimize a fuse-blowing scheme:

Use fast fuses or current-limiting fuses: If fuse blowing is used, no need exists to use big, slow fuses. If smaller or faster fuses are used, the fault will be cleared faster, so the voltagesag duration will be shorter. Current-limiting fuses can also be used to limit the magnitude and duration of the sag. Note that fuses should not be too small, or nuisance fuse operations will occur because of loading, inrush, and cold load pickup. Note that if current-limiting fuses are used, it would be difficult to switch back to a fuse-saving scheme. Covered wire or small wire: Watch for burndowns on circuits with covered wire or small wire that is protected by the station breaker or recloser. If either of these cases exists, then it is recommended that a modified fuse-blowing scheme with a time-delayed, instantaneous element be used. Use single-phase reclosers on longer laterals: A good way of maintaining some of the reliability of a fuse-saving scheme is to use single-phase reclosers instead of fuses on longer taps. Then temporary faults on these laterals will not cause permanent interruptions to customers on these laterals. More fuses: Another option is to add more second- and third-level fuses to laterals to further segment the circuit. Track lateral operations: Temporary faults on fused laterals will cause sustained interruptions. To minimize the impacts on lateral customers, it is beneficial if interruptions are tracked by lateral. Poorly performing laterals can be identified. Poorly performing sections can be patrolled, trees trimmed, animal guards installed, and other similar actions.

3-33

Protection Strategies

Time Delay on the Instantaneous Element (Fuse Blowing) An alternative implementation of a fuse-blowing scheme is to use a time delay on the instantaneous trip (rather than removing it)40. This practice has the advantage that faults do not last as long as they would if the relay went to a time-delay element. This reduces the chance of wire burndowns and reduces the duration of voltage sags for faults on the mains. A common delay is 0.1 seconds. An example implementation is shown in Figure 3-15, where a 0.1-second delay is added to the instantaneous operation. A 100-K fuse link is also shown in Figure 3-15. For the 100-K link, the scheme is actually a mixture of fuse saving and fuse blowing. For fault currents greater than roughly 1700 A, it is a fuse-blowing scheme (the fuse will clear before the instantaneous relay operates). For currents less than 1000 A, the scheme is a fuse-saving scheme (the breaker will trip before the fuse is damaged). Between 1000 and 1700 A, one or both devices may operate. Another option that is sometimes used with this scheme is to use a high-set instantaneous along with the delayed instantaneous. The high-set instantaneous would have no delay and be set to clear faults close to the station. This will remove the most damaging faults quickly (because they are the most likely to cause damage and cause the most severe voltage sags).

N.G. Engelman, Relaying Changes Improve Distribution Power Quality, Transmission and Distribution, pp. 72 76, May 1990.

40

3-34

Protection Strategies
100 K Fuse 1

Breaker clear time 5-cycle breaker + 0.1-sec relay time

Time [sec]

0.1

Delay on the instantaneous Fuse saving for currents below this value Fuse blowing for currents above this value

Fuse Damage Curve

0.01 100

1000 Current [A]

10000

Figure 3-15 Example of a Delayed Instantaneous Element Used for Fuse Blowing

Using a delay is a better fuse-blowing scheme than just removing the instantaneous. The disadvantage, and the reason that it may not be implemented as much, is that it is usually more difficult and costly to implement. For electromechanical relays, another timer relay must be added, and the relay scheme must be engineered. Many digital relays have this option available, so in that case it would be much easier to implement. High-Low Combination Scheme Another option is to use fuse blowing at the substation and fuse saving at downstream reclosers. The rationales for these choices are:

Fuse blowing at the substation: Fault currents are high near the substation, so it is difficult to get fuse saving to work there.

3-35

Protection Strategies

Fuse saving at downstream reclosers: Fault currents are lower downstream, and reclosers are faster, so fuse saving should work well there.

This scheme is easy to implement. The station instantaneous trip would be eliminated. Reclosers would be operated with a fast trip (that is, the A curve)most are already in this mode, so no changes would be necessary here. Control of the Protection Scheme by SCADA System Another option is to use a supervisory control and data-acquisition (SCADA) system to change back and forth between fuse saving and fuse blowing. This is an attempt to get some of the benefits of both schemes. Fuse blowing would be the normal operating mode, but human operators could switch to a fuse-saving scheme during storms. This would avoid clear-sky momentaries while at the same time improve storm restoration. Several factors make it better to use a fuse-saving scheme during storms:

Faults are more likely to be temporary during storms (lightning, wind). Customers are more forgiving about momentaries during storms. Outages caused by fuse operations last longer during storms (because crews have many repairs to perform). If fuses are blown because of temporary faults, this will increase the number of repair locations. Fuse saving will reduce the number of outages that crews will have to address.

For SCADA control of fuse saving to be effective, the system will have to be designed for fuse saving. This means using larger, slower fuses for laterals close to the substation, using faster breakers (or use substation reclosers), or possibly even using grounding reactors in the substation to limit fault currents. Likewise, the system would have to be designed for fuse blowing. This would mean avoiding using tree wire (or go to a delayed instantaneous rather than removing the instantaneous trip). Control is more readily available in the substation because the SCADA infrastructure may already be in place. If so, the cost of the SCADA system has already been justified, and this added functionality could be piggybacked on the existing system if there are free channels available. It is feasible to use automation technology to implement remote control of feeder reclosers, but the cost of the communication equipment probably would not justify having this functionality. For SCADA control, microprocessor-controlled relays are not needed. A SCADA channel can be used to control a blocking relay on the instantaneous elements of the feeder relays. An alternative would be for the SCADA channel to control the delay on the instantaneous elementOption 1 is no delay (or fuse saving); Option 2 has a delay (fuse saving below a certain fault current); and Option 3 is inhibited instantaneous operation (or fuse blowing). One SCADA channel could be used to control the fuse saving/blowing status of all of the distribution feeders in a station. Alternatively, each feeder could be controlled independently.

3-36

Protection Strategies

Adaptive Control by Phases Various protection schemes are classified as adaptive. An adaptive approach to a fuse-blowing mode is to adjust the scheme depending on how many phases are faulted:

Two- or three-phase fault: Use the instantaneousthe fault is assumed to be on the threephase mains. By tripping quickly, this will reduce the duration of voltage sags for faults on the mains. Single-phase fault: Use fuse blowing (that is, use time-delay curves or a delayed instantaneous relay).

This scheme would require microprocessor-based relays. This is not a common scheme, and the expense and complexity is not likely to justify it unless the chosen relays come with this functionality, and it is easy to configure the relay for this operation mode. Economics The hardware costs associated with manually changing back and forth between fuse saving and fuse blowing is relatively low. The majority of the cost is in the labor to re-fuse laterals and adjust relay settings. The major cost impact of using a fuse-blowing scheme is on restoration efforts. Fuse blowing will result in more trips by crews to replace blown fuses. Summary Fuse saving has traditionally been used for two reasons: fewer blown fuses to replace and improved reliability. The disadvantage of fuse saving is increased momentary outages for customers. When momentary outages become a major concern, particularly for customers connected to the main line, fuse blowing may make more sense. Fuse blowing will protect customers from momentary outages caused by faults on laterals. When implementing fuse blowing, care must be taken to avoid situations that could result in conductor damage or burndown. One popular scheme is to use fuse blowing for laterals between the station and the first downstream recloser and fuse saving for the remainder of the circuit.

3-37

Protection Strategies

Current-Limiting Fuses
Current-limiting fuses (CLFs) are fuses that have the ability to reduce the magnitude of fault current caused by downstream faults. CLFs are traditionally used in high-fault-current areas to reduce fault duty on equipment. More recently, they are being used to reduce the severity of voltage sags. The following benefits, negatives, economics, and bottom line are realized with the use of CLFs. Benefits

Major reduction in the depth of voltage sags and their duration compared to depth and sag when using an expulsion fuse An increase in safety for workers and the public because of greatly reduced damage at the fault location

Negatives

High cost compared to the cost of an expulsion fuse Incompatibility between fuse saving and CLF usage on lateral lines Susceptibility of CLFs to damage by transients An increased difficulty in coordinating with other protective devices

Economics CLFs are more expensive than standard fuses. One product by S&C Electric Company, the Fault Tamer, costs approximately four times more than a standard fuse. Many CLFs require special hardware for installationfurther increasing costs. While these costs are high compared to traditional fusing equipment, they may be justifiable if their use can improve the power quality of a customer with power quality sensitivity. Bottom Line CLFs can be useful in improving the power quality of a distribution circuit. They have a particularly positive impact on the depth of voltage sags caused by line-to-ground faults downstream of fuses. However, the improvement in power quality comes with an increase in hardware costs. Introduction CLFs consist of fuseable elements in silicon sand. When a fault current melts the fuseable elements, the sand will melt and create a narrow tube of glass called a fulgurite. The voltage

3-38

Protection Strategies

across the arc in the fulgurite will greatly increase, which causes the effective resistance of the arc to significantly increase. The constriction of the arc by the fulgurite causes high pressure around the arc, which creates conditions of high voltage across the arc. This back voltage reduces the current. The increase in resistance also lowers the X/R ratio, which causes a premature current zero. At the current zero, the arc extinguishes. Current-limiting fuses are noted for their extremely high fault-clearing capability. CLFs are available with symmetrical, maximum interrupt ratings up to 50 kA. Contrast that with expulsion fuses, which may have typical maximum interrupt ratings of 3.5 kA in oil and 13 kA in a cutout. Current-limiting fuses also completely contain the arc during operation and are noiseless with no pressure buildup. Current-limiting fuses cause a reduction in fault current and an early zero crossing. In the process, both the voltage-sag magnitude and duration are reduced. Figure 3-16 shows an 41 example of a fault cleared by a CLF . The duration of the sag is extremely short. These types of results have been verified by other measurements and computer models42,43.
Wave Fault 150 100 50
% Volts

0 -50 -100 -150 0 20

Wave Fault 40 20 0
Amps

Time (mSeconds)

40

60

80

100

-20 -40 -60 0 20

Time (mSeconds)

40

60

80

100

Figure 3-16 Voltage and Current Measured During Operation of a Current-Limiting Fuse in the EPRI Distribution Power Quality Study An Assessment of Distribution System Power Quality: Volume 3: Library of Distribution System Power Quality Monitoring Case Studies, EPRI, Palo Alto, CA: 1996. TR-106294-V3. 42 Kojovic, L.A., Hassler, S.P., Leix, K.L., Williams, C.W., and E.E. Baker, Comparative Analysis of Expulsion and Current-Limiting Fuse Operation in Distribution Systems for Improved Power Quality and Protection, IEEE Transactions on Power Delivery, Vol. 13, No. 3, pp. 863869, July 1998. 43 An Assessment of Distribution System Power Quality: Volume 3: Library of Distribution System Power Quality Monitoring Case Studies, EPRI, Palo Alto, CA: 1996. TR-106294-V3.
41

3-39

Protection Strategies

Industry Usage Current-limiting fuses are widely used for protection of equipment in areas with high fault current. Table 3-14 shows the percentages of utilities that use CLFs from a 1995 Institute of 44 Electrical and Electronics Engineers (IEEE) survey . The major reason given for the use of CLFs is safety; the second, most-common reason is high fault currents in excess of expulsionfuse ratings. (Effects on voltage sags were probably not a consideration in the survey.) Use on laterals was low and typically concentrated on underground laterals where fault currents are generally higher.

Working Group on Distribution Protection, W. M. Strang, Chair, Distribution Line Protection Practices Industry Survey Results IEEE Power System Relaying Committee Report, IEEE Transactions on Power Delivery, Vol. 10, No. 1, pp. 176186, Jan. 1995.

44

3-40

Protection Strategies Table 3-14 Percent of Utilities Using Different Types of Current-Limiting Fuses as Reported in a 1995 Institute of Electrical and Electronics Engineers Survey Uses of Current-Limiting Fuses General Purpose Backup On OH line laterals On UG line laterals 5 kV 15% 15% 5% 7% Line Voltage Class 15 kV 29% 38% 6% 18% 25 kV 30% 43% 9% 20% 35 kV 18% 30% 3% 18%

More recently, utilities have started to use CLFs for power quality reasons (or maybe as an extra justification)45. Types of Current-Limiting Fuses Current-limiting fuses can be classified into three basic types46:

Backup: A fuse capable of interrupting all currents from the maximum-rated interrupting current down to the minimum-rated interrupting current General purpose: A fuse capable of interrupting all currents from the maximum-rated interrupting current down to the current that causes melting of the fusible element in 1 hour Full range: A fuse capable of interrupting all currents from the rated interrupting current down to the minimum continuous current that causes melting of the fusible element(s), with the fuse applied at the maximum ambient temperature specified by the manufacturer

Current-limiting fuses are extremely good at clearing high-current faults. They have a much more difficult time clearing low-current faults or overloads. For a low-current fault, the fuseable element will not melt, but it will get extremely hot and can melt the fuse hardware, resulting in failure. The inability to handle low-current faults is why the most common application of CLFs is as a backup for an expulsion fuse. The expulsion fuse clears low-current faults, and the CLF clears high-current faults. General-purpose fuses usually use two elements in seriesone for the high-current faults and one for the low-current faults. General-purpose fuses could fail for overloads, so their application should be restricted to situations where overloads are not present or are protected by some other device (such as a secondary breaker on a transformer).

C. Price, Current-Limiting Fuses Enhance Power Quality, Transmission & Distribution World, Nov. 1998. IEEE Std. C37.40-1993, IEEE Standard Service Conditions and Definitions for High-Voltage Fuses, Distribution Enclosed Single-Pole Air Switches, Fuse Disconnecting Switches, and Accessories.
46

45

3-41

Protection Strategies

Full-range fuses provide even better low-current capability and can handle overloads and lowcurrent faults without failing (as long as the temperature is within rating). One consideration is that short-duration currents (lightning, inrush, or faults cleared by another device) can damage the fuse element in the full-range fuse. If the fuse element is damaged, localized heating and arcing can occur when carrying load current after the transient. The localized heating and arcing can cause failure of the fuse. Contrast this to an expulsion fuse. If a transient damages the expulsion fuse, the arcing in the fuse element will melt it completely. The fuse will operate, but the fuse holder will not be structurally damaged and no fault will be created on the system. Current-limiting fuses can be applied in several ways, including:

Backup CLF in series with an expulsion fuse in a cutout Full-range CLF in a cutout Backup CLF under oil Full-range (or general-purpose) CLF under oil CLF in dry-well canister or insulator

The typical application of CLFs for power quality purposes is installation on lateral taps. This means it will be a backup CLF such as AB Chance K-mate in series with an expulsion fuse, or it will be a general-purpose, drop-out CLF such as Cooper Power Systems X-Limiter. Safety Advantages Current-limiting fuses have the safety advantage of much lower energy at the location of the fault. This lower energy provides safety to workers and the public. Damage to life and property can occur in several ways:

Pressure wave: The fault arc pressure wave can cause damage to equipment and personnel. Heat: The heat from the fault arc can cause burns to personnel and start fires. Pressure buildup: A fault arc in oil can cause pressure buildup that can rupture equipment.

All of these effects are related to the arc energy. The arc energy is primarily a function of It (total charge, in coulombs).* The energy at the damage point is a function of the arc voltage times arc current. The arc voltage stays relatively constant at a maximum of approximately 200 to 250 V/in. over a wide current range (the voltage along an arc in oil is higher than along an arc in free air). Distribution transformers can fail violently because of internal failures. Current-limiting fuses can greatly reduce the chance of a transformer failing violently. The pressure-release valves on

It is a common misconception that fault damage is a function of I2t. I2t only applies if the resistance stays constant (as in a wire), but in an arc under oil, the voltage stays constant at approximately 200 to 250 V/in., and the resistance varies.

3-42

Protection Strategies

transformers are not fast enough to significantly reduce the pressure buildup during high-current faults.
47 According to industry standards , distribution transformers with external fuses are subjected to a test where an internal arcing fault with an arc length of 1 inch is maintained for to 1 cycle. The current is 8000 A. Under this fault condition, the transformer must not be ruptured or expel excessive oil. For internally fused, completely self-protected (CSP), or pad-mounted transformers, the arc test is performed to the rating of the fuse, which is generally much lower than 8000 A.

Table 3-15 shows a summary of the maximum currents used to test different types of transformers48. The fuse in a CSP transformer is the reason for the lower rating. When the underoil fuse melts, it will create another location for an arc in the transformer (in addition to the fault). The reason that higher-voltage CSP transformers have a lower rating is that the fuse is longer (approximately 3 in. for a 15-kV fuse and 5 in. for a 35-kV class fuse). The longer the fuse, the longer the arc after the fuse element melts. With a longer arc, the energy in the transformer will be larger. Note that this test does not include all of the possible failure modes, and there is no guarantee that a transformer will not fail with lower current. For example, a failure with an arc longer than 1 inch will have more energy and could rupture the transformer at a lower level of current.
Table 3-15 Maximum - to 1-Cycle Current Rating on Distribution Transformers Maximum Tested Symmetrical Current (A) Overhead transformer Under-oil expulsion fuse (up to 8.3 kVLG) Under-oil expulsion fuse (up to 14.4 kVLG) Under-oil expulsion fuse (up to 25 kVLG) 8000 3500* 2500* 1000*

Transformer

*Based on typical under-oil ratings of the fuse links. If a transformer is applied in a location where the available line-to-ground fault current is higher than shown in Table 3-16, then it is recommended that CLFs be used. CLFs can provide improved transformer protection when installed at the transformer or upstream at the lateral tap point. As a comparison, an 80-A, 8.3-kV Cooper ELF (a current-limiting fuse that might be installed on a lateral) has a maximum I2t rating of 150,000 A2s. A 25-A fuse of the same type has a
ANSI Std. C57.12.20-1988, Overhead-Type Distribution Transformers, 500 kVA and Smaller: High Voltage, 34500 Volts and Below; Low Voltage 7970/13800Y and Below. 48 ANSI Std. C57.12.20-1988, Overhead-Type Distribution Transformers, 500 kVA and Smaller: High Voltage, 34500 Volts and Below; Low Voltage 7970/13800Y and Below.
47

3-43

Protection Strategies

maximum I2t rating of 1700 A2s. Typical I2t ratings of common fuses are shown in Table 3-17. 2 2 With no CLF, a 10-kA fault for a -cycle will give an I t of 833,333 A s. To relate these 49,50* numbers to what a transformer can withstand, see Table 4 .
Table 3-16 Maximum I2t Values for Common Current-Limiting Fuses Voltage Rating (kV) 8.3 8.3 8.3 8.3 8.3 8.3 Fuse Rating (A) 12 25 80 12 25 80 Maximum I2t (A2s) 7000 17,000 150,000 18,000 45,000 720,000

Fuse Cooper ELF Cooper ELF Cooper ELF Chance K-Mate Chance K-Mate Chance K-Mate

Table 3-17 Typical Transformer I2t Capabilities Transformer 15-kV Overhead 25-kV Overhead 35-kV Overhead 15-kV Underground 25-kV Underground 35-kV Underground I2t Capability (A2s) 120,000 66,000 50,000 500,000 300,000 100,000

Best Locations for Use The best location for use of CLFs on distribution systems is close to the station. This location is where they are most effective in limiting damage from high fault currents. This location is also where they are most effective for reducing the magnitude and duration of a voltage sag. Refer to
Walgren, C., Leix, K. and J.H. Webb, Current Limiting Fuse Addresses Safety and Power Quality Concerns, IEEE Rural Electric Power Conference, 1992. 50 A.C. Westrom, Distribution Fusing Applications, International Conference of Distribution Fusing, 1981.
* 49

Even though transformer damage is not really a function of I2t, it is the most cited parameter from test data. It is probably most valid for faults of approximately 8 kA for cycle.

3-44

Protection Strategies

Figure 3-17 for voltage-sag magnitudes as a function of distance from the station and refer to Figure 3-18 for fault currents. At a 12.5-kV line voltage, current-limiting fuses used in the first 1 to 2 miles will reduce the depth of voltage sags on the station bus and reduce the impact of fault currents (those greater than 2 to 4 kA).

0 .8 P e r u n it lin e -to -g ro u n d vo lta g e

12kV 25kV

0 .6

0 .4

0 .2

Single-phas e f ault Three-phas e f ault

0 .0 0 2 4 6 8 D ista nce fro m the sub sta tio n, m ile s

Figure 3-17 Voltage Sags on the Station Bus Versus Fault Location as a Function of Distance From the Station

3-45

Protection Strategies

10

8 F a u lt cu rre n t, kA

Three-phas e f ault Single-phas e f ault

25kV 12kV

D ista nce fro m the sub sta tio n, m ile s

Figure 3-18 Fault Current Versus Fault Location as a Function of Distance From the Station

For critical customers that are not close to the station, faults beyond the first 1 to 2 miles will also cause severe sags. In these situations, using CLFs on lateral taps beyond the first 1-2 miles will help only on the feeder serving the critical customer. CLFs only help on adjacent feeders when installed close to the station. The benefit will not be as great as it is for CLFs close to the station (the current-limiting effect is stronger for higher fault currents). Current-limiting fuses are normally used right on the equipment being protected (usually transformers). For power quality improvement, however, use on lateral taps is most effective. Most faults on overhead lines occur on the line itselfnot because of transformer failures. The same is true for cables. CLFs applied at lateral taps will still provide some protection for transformers. As a comparison, an 80-A, 8.3-kV Cooper ELF current-limiting fuse, which 2 2 might be applied to a lateral tap, has a maximum I t rating of 150,000 A s. A 25-A fuse of the same type has a maximum I2t rating of 1700 A2s. With no CLF, a 10-kA fault for a -cycle will give an I2t of 833,333 A2s. This shows that both the 80-A and 25-A fuses protect the transformer to I2t values less than shown in Table 3-17 (the case without the CLF for a 10-kA fault exceeding the transformer capabilities). At higher voltages or larger lateral fuse sizes, a CLF at the tap point may have a maximum I2t value that exceeds transformer withstand capability. Another consideration is that lateral CLFs may not be available for higher-voltage systems. CLFs are available for 15-kV class systems in sizes up to 100 A.

3-46

Protection Strategies

Drawbacks Some of the drawbacks of CLFs include:

Voltage transient from CLF operation: When a CLF operates, the rapidly changing current will cause a voltage spike (V = Ldi/dt). Usually, this voltage transient is not severe enough to cause problems for the fuse or for customer equipment. Figure 3-16 shows a mild voltage transient with the fuse operation. Limited overload capability: A backup or general-purpose fuse does not do well for overloads or low-current faults. A full-range fuse performs better, but it still could have problems with a transient overcurrent that partially melts the fuse. Coordination issues: A current-limiting fuse may be difficult to coordinate with other feederprotective devices. CLFs are fast enough that they almost have to be used in a fuse-blowing scheme (fuse saving will not work because the fuse will be faster than the breaker). Also, it may be difficult to coordinate a lateral CLF with a downstream expulsion fuse. High cost relative to an expulsion fuse.

Economics Current-limiting fuses can be from four to eight times as expensive as expulsion fuses. This cost comes from both the increased cost of the device itself and, in some cases, increased cost in mounting hardware. While these costs are high compared to traditional fusing equipment, they may be justifiable if their use can positively affect the power quality of a sensitive customer. Summary CLFs have historically been used to protect primary equipment, particularly transformers, in areas with high fault current availability. In recent years, the ability of a CLF to reduce the depth of voltage sags caused by line-to-ground faults has led to their use to fuse laterals. This increased use is particularly true for laterals between the circuit breaker and the first downstream recloser. This section of line typically has the highest available fault currents. The use of CLFs in these sections of the circuit can provide an improvement in power quality to customers served from the remainder of the feeder.

Sequence Coordination of Reclosers

Sequence coordination is a scheme used to inhibit the tripping of a recloser (or breaker) on its fast curve when a downstream recloser operates on a fault. The use of sequence coordination can greatly reduce the number of momentary outages experienced by distribution customers. Benefits

Reduces the number of momentary outages for distribution customers.

3-47

Protection Strategies

Reduces customer complaints. Is not expensive and can be quickly implemented on circuits where electronically controlled reclosers and microprocessor-based relays are used.

Negatives

Is difficult to implement in areas with high fault currents. Depends on an upstream device being either an electronically controlled recloser or a breaker controlled by a microprocessor-based relay.

Economics Sequence coordination is not expensive to install for circuits with the required protection equipment. Introduction Most utility reclosers (or breakers) are programmed to open once (or twice) using a fast curve (known as a fast operation) and then open once (or twice) using a slower curve (known as a delayed operation) during a permanent fault. The most common setup is for two fast operations and two delayed operations before the recloser locks out. The recloser thus opens three times allowing the fault to clear itselfand then opens a fourth time and locks out if the fault is still present. This system is meant to reduce the number of permanent outages from temporary faults. With two or more reclosers (or breakers) in series, it is possible for more than one device to respond to a given fault. This creates momentary interruptions on sections of the distribution feeder that are not faulted, thus degrading the quality of power provided to those customers. Sequence coordination inhibits the fast tripping of all devices except the one immediately upstream from the fault, reducing the number of momentary interruptions experienced by customers on the feeder. Standard Coordination and Its Effect on Momentary Interruptions Standard coordination methods for reclosers can result in unnecessary momentary interruptions for customers not on the faulted zone of a circuit. Figure 3-19 illustrates what can happen for a faulted circuit with two reclosers, both with two fast and two slow operations. Recloser R2 (the device immediately upstream from the fault) is supposed to operate for faults in its zone. However, if the fault current is high enough, Recloser R1 (the next device upstream of Recloser R2) may also operate. This is because the fast curve of the upstream recloser is often faster than the slow curve of the downstream recloser.

3-48

Protection Strategies

R1 Source

R2 Fault

Closed

R2
Open Without Sequence Coordination Closed

R1
Open Closed

R2
Open With Sequence Coordination Closed

R1
Open Time

Fault Start

Figure 3-19 Performance of Reclosers With and Without Sequence Coordination

When the fault occurs, the closest upstream device (Recloser R2 in this case) operates on its fast curve. In this example, the fault is permanent. Therefore, R2 will reclose and then operate again on its fast curve. As the illustration shows, the next upstream device will operate twice on its fast curve. This is because the fast curve for R1 is faster than the delayed curve of R2. Once R1 has operated twice on its fast curve, R2 will operate twice on its delayed curve and lock out. Sequence Coordination and Its Effect on Momentary Interruptions
51 Many utilities are now attempting to create a high-reliability zone for the section of the distribution feeder between the station and the first mainline protection device (that is, between R1 and R2 in Figure 3-19). High-reliability zones are created using a variety of techniques, including changing to a fuse blowing scheme, using instantaneous reclose, using current-limiting fuses on laterals, and using sequence coordination. This high-reliability zone is treated differently because an outage in this zone affects every customer on the circuit. Reducing the

B.D. Coate and D.C. Wareham, Decreasing Momentary Outages with Specific Distribution Feeder Improvements, IEEE Transactions on Industry Applications, Vol. 37, No. 2, March-April 2001, pp. 458463.

51

3-49

Protection Strategies

number of momentary and permanent outages in the high-reliability zone can greatly increase the reliability of the entire circuit. With sequence coordination, a protective devices fast curve is inhibited for a fault beyond the protection zone of the device if a downstream device opens first. In our example, Figure 3-19, the fault is outside of the protective zone of Recloser R1. The fault is in the protective zone of R2. With sequence coordination properly implemented, a permanent fault will result in the following sequence of operations:

R2 operates on its fast curve and recloses. R2 operates a second time on its fast curve and recloses. R1 inhibits its fast curves and does not operate. (But it does count the event and advance its control sequence.) R2 operates twice on its delayed curve and locks out. (R1 does not operate because it has advanced its control sequence to operate on its delayed curve.)

In this example, sequence coordination prevents two momentary operations for the highreliability zone of the circuit. With normal coordination, every customer on the circuit would have been subjected to two unnecessary momentary interruptions. Comparison of Reliability Using a simple circuit, it is possible to illustrate the difference in reliability performance in a circuit with sequence coordination. For this illustration, we will use a circuit with two protection zones, evenly spaced, and with equal number of customers in each zone (see Figure 3-20). For our calculations, we will assume that each zone experiences five permanent outages per year.

R1 Zone 1 Customers

R2 Zone 2 50 Customers

Figure 3-20 Example Distribution Circuit

Sequence coordination has the greatest effect on the Momentary Average Interruption Frequency Index (MAIFI). MAIFI represents the average number of times a customer is subjected to a momentary outage and is defined in Equation 3-7.

MAIFI =

ID N
i

NT

Eq. 3-7

3-50

Protection Strategies

where i = interruption event, IDi = number of interrupting device operations, Ni = number of interrupted customers for each interruption event, and NT = total number of customers served. As you can see in Table 3-18 sequence coordination can eliminate up to two operations for Recloser R1 for each fault in Zone 2 of the circuit. This has a significant impact on MAIFI because an operation of Recloser R1 affects all 100 customers of the circuit.
Table 3-18 Number of Device Operations for a Fault in the Specified Zone (Assuming Two Fast and Two Delayed Operations Before Lockout) Fault in Zone 1 Device Without Sequence Coordination 4 0 With Sequence Coordination 4 0 Fault in Zone 2 Without Sequence Coordination 2 4 With Sequence Coordination 0 4

Recloser R1 Recloser R2

The MAIFI for this circuit (see Figure 3-21) without sequence coordination, assuming five permanent faults per year per zone, is calculated as follows: First, calculate IDiNi for faults occurring in each zone: Fault in Zone 1: (4 operations 100 customers for R1) + (0 operations 50 customers for R2) = 400 per event This is multiplied by the number of events (five in this case) for a total of 2000. Fault in Zone 2: (2 100) + (4 50) = 400 400 5 = 2000

Now the IDiNi terms are added together and divided by the total number of customers served: Sum of customer outage numbers = 2000 + 2000 = 4000 MAIFI = 4000/100 = 40 interruptions per year per customer

Similarly, the MAIFI for the circuit with sequence coordination is calculated as follows:

3-51

Protection Strategies

Faults in Zone 1: Faults in Zone 2:

[(4 100) + (0 50)] 5 = 2000 [(0 100) + (4 50)] 5 = 1000

MAIFI = (2000 + 1000)/100 = 30 interruptions per year per customer Using sequence coordination reduced the MAIFI for this circuit by 25%. Granted this is a simplistic example. However, the use of sequence coordination has been shown to significantly reduce MAIFI numbers and customer complaints in utilities. Coordination Requirements For sequence coordination to work, the fast curves of the two devices must be properly coordinated. Again referring to the example circuit in Figure 3-21, Reclosers R1 and R2 must be properly coordinated. The control response curve of Recloser R1 must be slower than the recloser clear curve of R2 (that is, the control response curve of R1 must be above the recloser clear curve of R2). The control response curve is the curve representing the time it takes for the recloser control (or relay) to determine that there is a fault condition and send a trip signal to the recloser (or breaker). The total clear curve is simply the control response curve plus the amount of time it takes for the recloser (or breaker) to open once it receives a trip signal. Figure 3-22 illustrates the coordination of the two reclosers in our example circuit, assuming R1 has a trip setting of 200 amps and R2 has a trip setting of 100 amps. The R1 curve is the control response curve. The R2 curve is the recloser clear curve. As you can see, the devices only coordinate up to approximately 1500 amps of fault current. For any fault greater than this, sequence coordination will not work properly and R1 will also open for faults in Zone 2.

3-52

Protection Strategies
1

Time (seconds)

0.1

R2 R1

0.01 10 100 1000 Current 10000 100000

Figure 3-21 Coordination of the Fast Curves of Two Reclosers

There are many options for correcting this problem if necessary. Most recloser controls and relays offer several curves to choose from. Many also allow the operator to modify the curve to provide better coordination in this type of situation. One such option is minimum response. Setting a minimum response time forces the control to wait at least that amount of time before sending a trip signal to the recloser or breaker. A minimum response setting of 0.06 seconds is illustrated in Figure 3-23. This setting would allow these reclosers to be coordinated for all fault currents, which would allow sequence coordination to work properly for all levels of fault current.

3-53

Protection Strategies
1

Time (seconds)

0.1

R1 R2

0.01 10 100 1000 10000 100000

Current

Figure 3-22 Coordination of the Fast Curves of Two Reclosers, One With Minimum Response Set to 0.06 Seconds

If the recloser (R2 in our example) is a hydraulic, single-phase recloser, both the phase and ground fast curves of the upstream recloser (R1 in our example) must coordinate with the downstream devices fast trip curve. More discussion of sequence coordination and the modification of trip curves can be found in Electrical Distribution-System Protection by Cooper Power Systems52. Summary Sequence coordination can improve the reliability and power quality of a distribution circuit by reducing the number of momentary outages experienced by customers on the circuit. Sequence coordination is extremely effective at reducing the number of momentary outages experienced on circuit zones that are not faulted. The only requirements are that the upstream device must be electronically controlled (either a recloser control or a relay) and the fast curves of the two devices must coordinate properly over the entire range of expected fault currents. Coordination of the fast curves can be difficult for high fault currents. However, most controls allow for modification of the response curves to provide the necessary separation to allow proper coordination.

52

Electrical Distribution-System Protection, 3rd ed., Cooper Power Systems, 1990.

3-54

Protection Strategies

Locating Sectionalizing Equipment

To locate and prioritize sectionalizing projects, we can estimate the number of customer interruptions avoided by the addition of the recloser, fuse, or other sectionalizing device. The main inputs for this calculation are fault rates and circuit lengths. Consider the common but simple example of applying a recloser at the midpoint of a circuit mainline in Figure 3-23. The recloser protects the customers on the upstream side of the recloser from faults on the mainline downstream of the recloser. To estimate the number of customer interruptions saved, multiply the number of customers affected (all of those customers upstream of the recloser) by the fault rate times the mainline exposure downstream of the recloser. If we were finding locations for reclosers, the number of customers saved by each possible recloser location allows us to prioritize possible recloser locations. When comparing different equipment options, the cost effectiveness of each application can be determined by using the ratio of the cost of the installation of the sectionalizing device to the number of customer interruptions saved. Then, select projects with the lowest cost per customer interruption saved. On radial circuits with a major tee, the tee point is a prime candidate for a sectionalizing device on both branches off of the tee point. Figure 3-23 shows an example comparing recloser options at a tee point, the most number of customer interruptions saved is for reclosers on both branches off of the tee point. The best device (recloser or fuse) and the best location depend on the lengths, customer counts, and customer locations.

3-55

Protection Strategies

(a) Midpoint recloser Fault rate = 0.2 faults/mile/year 3-phase recloser installation = $15,000
Source

R
400 customers 3 miles Customer-interruptions saved = 400 customers (3 miles) (0.2 faults/mile) = 240 customer interruptions Cost per customer interruption saved $15, 000 = = $62.5 / cust. int. saved 240 cust. int. saved (b) Tee point reclosers 300 customers 400 customers, 3 mi.

B C
400 customers, 3 mi.

Option A: 300 cust (6 mi) (0.2 faults/mi) = 360 cust. int. saved Option B or C: 700 cust (3 mi) (0.2 faults/mi) = 420 cust. int. saved Option C with A already there: 400 cust (3 mi) (0.2 faults/mi) = 240 cust. int. saved Option C with B already there: 700 cust (3 mi) (0.2 faults/mi) = 420 cust. int. saved Costs per customer interruption saved: Option A: $15,000 / 360 = $ 41.7 / cust. int. saved Option B or C: $15,000 / 420 = $ 35.7 / cust. int. saved Option A and C: $30,000 / (360 + 240) = $ 50.0 / cust. int. saved Option B and C: $30,000 / (420 + 420) = $ 35.7 / cust. int. saved

Figure 3-23 Recloser Examples with Customer Interruptions Saved

Figure 3-24 shows that applying a tap fuse is not surprisingly a very cost-effective way to reduce the number of customer interruptions. For longer taps, we may consider using a single-phase recloser rather than a fuse, and this question calls for a somewhat different analysis approach. One point to keep in mind is that most applications of sectionalizing equipment will not dramatically reduce utility restoration costs; it may take less time to patrol and find the damage, but the largest costs associated with restorationgetting the crews to the site and the repair timedo not change. On overhead circuits, a recloser instead of a fuse will reduce the number of outage events by clearing temporary faults as a momentary interruption rather than having a blown fuse result in a sustained interruption. Consider a 2.5-mi tap with 50 customers. A fuse on that tap might blow (2.5 mi) (0.4 faults/mi/year) = 1 operation per year (assuming a fuse-blowing mode where the fuse operates for both temporary and permanent faults). Having a recloser on the tap may reduce the number of outage events on the tap by a factor of two (one operation every 3-56

Protection Strategies

two years). If a single-phase recloser installation costs $1500 more than a fuse installation, and each fuse operation costs $800 on average to replace, the recloser will pay for itself in four years. In this case, it is easier to estimate the impact by event, rather than using a cost per customer interrupted.

Fault rate = 0.2 faults/mile/year 1-phase fuse installation = $400 0.3 miles

1000 customers Customer-interruptions saved = 1000 customers (0.3 miles) (0.2 faults/mile) = 60 customer interruptions Cost per customer interruption saved $400 = $6.7 / cust. int. saved = 60 cust. int. saved

Figure 3-24 Tap Fuse Example with Customer Interruptions Saved

To apply sectionalizing equipment, Duke Power uses the customer-interruptions-saved approach. They will implement those projects where the cost is less than $50 per customer interruption saved using the cost guidelines in Table 3-19. These estimates are approximate and are intended to give engineers a feel for what is justified in sectionalization work. When protecting customers upstream from faults downstream, Duke Power assumes an annual permanent fault rate on overhead circuits of 0.2 faults/mile (0.32 faults/km) for circuit mainlines and 0.5 faults/mile (0.8 faults/km) for branch lines. Duke also considers a temporary fault rate that will manifest as a sustained interruption unless you use a reclosing device or fuse saving. This temporary fault rate is needed to analyze the impact of a sectionalizing device to protect customers downstream from faults downstream, such as when considering the impact of fuse saving or considering whether a recloser is more appropriate than a fuse. For either mainlines or branch lines, they assume a temporary fault rate of 0.2 for retrofitted lines (those where transformer locations have been retrofitted with a local fuse along with other upgrades to reduce faults) and 0.8 for unretrofitted lines. To estimate the impact of fuse saving using these temporary fault rates, consider the impact of each fuse downstream of the reclosing device. If the fuse coordinates with the upstream device (meaning the breaker or recloser can open before the fuse blows), then multiply the length of that tap by the temporary fault rate on that tap to arrive at the number of customer-interruptions saved with fuse saving. For a 1/2-mile long tap with 20 customers and a temporary fault rate of 0.8/mile/year, fuse saving will save 8 customer interruptions annually on that tap (add up the remaining taps to evaluate the full impact). Using Duke Powers numbers, the value of each sectionalizing device can be estimated by multiplying the fault rate by the value of each customer interruption ($50) by the exposure, or one can multiply the line exposure and the number of customers affected and then multiply by $10/customer-mile ($50/customer interruption 0.2 faults per mile). Consider the example in 3-57

Protection Strategies

Figure 3-25 when comparing two recloser scenarios. In this case, Duke can justify installing three single-phase reclosers at either location because both locations have a value of $9000, which is less than the installation cost of $7500. Installing reclosers at both locations is not justified as the value of the second recloser location is only $3000. And it is not beneficial to move the reclosers from location B to A (or vice versa) because both have the same number of customer-interruptions saved.
Table 3-19 Duke Power Sectionalizing Cost Guidelines Activity Install one fuse Install three fuses Install three single-phase hydraulic reclosers Install a three-phase vacuum recloser Install a three-phase electronic recloser Relocate a recloser Approximate Cost $200 $500 $7500 $9500 $20000 $3750

Fault rate = 0.2 faults/mile/year

Source

A
300 customers 150 customers 1 mile

B
550 customers 2 miles

Customer exposure value = (0.2 faults/mi)($50/customer-interruption saved) = $10/customer-mile Value of each option: A only: 300 cust (3 mi) ($10/customer-mile) = $9000 B only: 450 cust (2 mi) ($10/customer-mile) = $9000 A with B already there: 300 cust (1 mi) ($10/customer-mile) = $3000 B with A already there: 150 cust (2 mi) ($10/customer-mile) = $3000

Figure 3-25 Duke Power Sectionalizing Example

Customer groupings are important to consider. Sometimes it makes sense to increase the mainline exposure to protect large customer groupings. For example if there is a large grouping of customers near the middle of a circuit, putting the recloser just beyond that grouping will improve reliability the most. On circuits with large groupings of customers near the end will have few opportunities for additional sectionalizing equipment unless automated tie-point equipment is used to transfer those customers to another circuit. For long straight radial circuits, one recloser may be beneficial, but having a second recloser may be difficult to justify (although an automated tie recloser may be very beneficial). 3-58

Protection Strategies

For more advanced analysis, modules for reliability evaluations are available for most distribution analysis software, and some offer automatic optimization routines to help apply sectionalizing equipment. Distribution reliability software allows more fine-tuned estimates that can handle different fault rates and account for temporary and permanent faults, possibly account for the coordination between devices, and possibly account for the differences between singlephase and three-phase devices.

3-59

4
CONSTRUCTION ISSUES
Robust line construction plays a big role in maintaining reliable service. On overhead distribution and subtransmission circuits, many faults result from inadequate clearances, inadequate insulation, or old equipment. Many of these faults are preventable with proper design, implementation, and maintenance practices. The best way to reduce faults over time is to "institutionalize" fault-reduction practices.

Fault-Reduction Methods and Processes

Because faults are the root cause of voltage sags and interruptions, reducing the number of faults will obviously reduce the number of voltage sags and interruptionsand power quality will improve. Faults are not evenly distributed along lines. Faults are not inevitable. Not all faults are acts of God. Most are from specific deficiencies at specific structures. On overhead distribution circuits, most faults result from inadequate clearances, inadequate insulation, old equipment, or from trees or branches falling into a line. Consider faults as preventable. Now, we need to consider programs to reduce faults. These programs need to address inadequacies in design standards and implementation practices. The best way to reduce faults over time is to institutionalize fault-reduction practices. After identifying the most common fault sources, implement programs to address these so performance improves continually. Start with a good design that eliminates fault sources, especially at equipment poles; use sufficient electrical clearance. Separate grounded objects from phase conductors as much as possible. Then employ procedures to ensure that more fault-resistant designs are implemented. Train linemen and field engineers to do it right. If possible, implement programs to bring old construction up to specifications; replace old arresters, increase poor clearances, add covered jumper wires, add animal guards, and so on. An opportune time to clean up poor construction is when crews are already doing work on a structure. Perform quality audits during work and after work is done. Give crews and field engineers feedback. Duke Power was instrumental in developing some of these techniques and programs to reduce faults (Chow and Taylor53). Quoting from Taylor54: 1. Every fault has a cause that can be identified and prevented.
Chow, M. Y., and Taylor, L. S., A Novel Approach for Distribution Fault Analysis, IEEE Transactions on Power Delivery, vol. 8, no. 4, pp. 18829, October 1993. 54 Taylor, L., Elimination of Faults on Overhead Distribution Systems, IEEE/PES Winter Power Meeting, 1995. Presentation to the Working Group on System Design, 1995 IEEE Winter Power Meeting, New York, NY.
53

4-1

Construction Issues

2. To stop the fault, you must find the exact cause and eliminate it. There are no magic programs or devices. The processes described here consist of several ongoing programs. Utilities must maintain consistency. For the most part, these are not one- or two-year programs. Maintaining fault resistance must be an ongoing process that becomes a core part of utility operations. Fault Basics Before tackling programs to reduce faults, we will consider some of the basic characteristics of faults. Knowing the causes and characteristics of faults is a good first step in eliminating as many of them as we can. There are many causes of faults on distribution circuits. A large EPRI study was done to characterize distribution faults in the 1980s at 13 utilities monitoring 50 feeders (Burke and Lawrence55 and EPRI 1209-156). The distribution of the causes of permanent faults found in the EPRI study is shown in Figure 4-1. Many of the fault causes are discussed in more detail in this chapter. Approximately 40% of faults in this study occurred during periods of adverse weather, which included rain, snow, and ice.
Lightning Tree contact Equipment failure Animal Wind Digin Vehicle accident Ice/snow Vandalism Construction activity Other 0 5 10 15 20 25

Percent of faults by cause

Data sources: Burke and Lawrence , EPRI 1209-1
55 56

Figure 4-1 Fault Causes Measured in the EPRI Fault Study

Distribution faults occur on one phase, on two phases, or on all three phases. Single-phase faults are the most common. Almost 80% of the faults measured involved only one phase either in contact with the neutral or with ground (see Table 4-1). As another data point, measurements on 34.5-kV feeders found that 75% of faults involved ground (also 54% were phase to ground, and

55

Burke, J. J. and Lawrence, D. J., Characteristics of Fault Currents on Distribution Systems, IEEE Transactions on Power Apparatus and Systems, vol. PAS-103, no. 1, pp. 16, January 1984. 56 Distribution Fault Current Analysis, EPRI, Palo Alto, CA: 1983. 1209-1.

4-2

Construction Issues

15% were phase to phase) (Johnston et al.57). Most faults are single phase because most of the overall length of distribution lines is single phase, so any fault on single-phase sections would only involve one phase. Also, on three-phase sections, many types of faults tend to occur from phase to ground. Equipment faults and animal faults tend to cause line-to-ground faults. Trees can also cause line-to-ground faults on three-phase structures, but line-to-line faults are more common. Lightning faults tend to be two or three phases to ground on three-phase structures.
Table 4-1 Number of Phases Involved in Each Fault Measured in the EPRI Fault Study Fault One phase to neutral Phase to phase Two phases to neutral Three phase One phase on the ground Two phases on the ground Three phases on the ground Other
58

Percentage 63% 11% 2% 2% 15% 2% 1% 4%

Sources: Burke and Lawrence , EPRI 1209-159

Figure 4-2 shows fault rates found in various studies for predominantly overhead circuits. Ninety faults per 100 miles per year (55 faults/100 km/year) is common for utilities with moderate lightning. Faults rates increase significantly in higher lightning areas. This type of data is difficult to obtain.

Johnston, L., Tweed, N. B., Ward, D. J., and Burke, J. J., An Analysis of Vepco's 34.5 kV Distribution Feeder Faults as Related to Through Fault Failures of Substation Transformers, IEEE Transactions on Power Apparatus and Systems, vol. PAS-97, no. 5, pp. 187684, 1978. 58 Burke, J. J. and Lawrence, D. J., Characteristics of Fault Currents on Distribution Systems, IEEE Transactions on Power Apparatus and Systems, vol. PAS-103, no. 1, pp. 16, January 1984. 59 Distribution Fault Current Analysis, EPRI, Palo Alto, CA: 1983. 1209-1.

57

4-3

Construction Issues

Southern US Florida South Carolina Southeastern US Virginia Ontario, Canada Northeastern US Ontario, Canada England Australia 0 47 35 19 99 95 90 89 70 317

351

High lightning, 90% overhead [Short, 2001] Very high lightning [Parrish, 1991] Relatively high lightning [Peele, 2001] Very high lightning [Short, 2001] Relatively high lightning, 34.5 kV [Johnston et al., 1978] Moderate lightning [CEA, 1998] Moderate lightning, 2/3 overhead [Short, 2001] Low lightning [CEA, 1998] Low lightning [Laverick, 1984] 11 kV [Littler, 1965]

100

200

300

400

Fault rate per 100 circuit miles per year

Figure 4-2 Fault Rates Found in Different Studies

Faults are either temporary or permanent. A permanent fault is one where permanent damage is done to the system. This includes insulator failures, broken wires, or failed equipment such as transformers or capacitors. Virtually all faults on underground equipment are permanent. Most equipment fails to a short circuit. Permanent faults on distribution circuits usually cause sustained interruptions for some customers. To clear the fault, a fuse, recloser, or circuit breaker must operate to interrupt the circuit. The most critical location is the three-phase mains, because a fault on the main feeder will cause an interruption to all customers on the circuit. A permanent fault also causes a voltage sag to customers on the feeder and on adjacent feeders. Permanent faults may cause momentary interruptions for a customer. A common example is a fault on a fused lateral (tap). With fuse saving (where an upstream circuit breaker or recloser attempts to open before the tap fuse blows), a permanent fault causes a momentary interruption for customers downstream of the circuit breaker or recloser. After the first attempt to save the fuse, if the fault is still there, the circuit breaker allows the fuse to clear the fault. If a fault is permanent, all customers on the circuit experience a momentary, and the customers on the fused lateral experience a sustained interruption. A temporary fault does not permanently damage any system equipmentif the circuit is interrupted and then reclosed after a delay, the system operates normally. Temporary (nondamage) faults make up 5090% of faults on overhead distribution systems. Some causes of faults are more likely than others to be permanent (see Table 4-2). The causes of temporary faults include lightning, conductors slapping together in the wind, animals that cause faults and fall off, and insulator flashovers caused by pollution. Temporary faults are the main reason that reclosing is used almost universally on distribution circuit breakers and reclosers (on overhead 4-4

Construction Issues

circuits). Temporary faults will cause voltage sags for customers on the circuit with the fault and possibly for customers on adjacent feeders. Temporary faults cause sustained interruptions if the fault is downstream of a fuse and fuse saving is not used or is not successful. For temporary faults on the feeder backbone, all customers on the circuit are momentarily interrupted. Faults that are normally temporary can turn into permanent faults. If the fault is allowed to remain too long, the fault arc can do permanent damage to conductors, insulators, or other hardware. In addition, the fault current flowing through equipment can do damage. The most common damage of this type is to connectors or circuit interrupters such as fuses.
Table 4-2 Normal Temporary or Permanent Fault Causes Temporary Faults Lightning Animals Wind Insulator contamination Permanent Faults Fallen trees or tree limbs Cable and accessory failures Other equipment failures Dig-ins

Many distribution faults involve arcs through the air, either directly through the air or across the surface of hardware. Although a relatively good conductor, the arc is a very hot, explosive fireball that can cause further damage at the fault location (including fires, wire burndowns, and equipment damage). Damage at the location is a function of the fault current magnitude and the duration of the fault. At faulted locations, the damage is sometimes obvious; an arc can melt a significant amount of metal, and scarring of equipment is apparent (see Figure 4-3 for an example). That said, some faults at some structures leave very little evidence that they occurred. Even some faults that cause permanent damage can be quite hard for crews to find.

4-5

Construction Issues

Charring due to power arc.

A portion of the metal base was vaporized by the power arc.

Figure 4-3 Power Arc Damage on the Base of a Fiberglass Insulator Standoff

Programs to Improve Construction Practices The main steps we will consider in a process to upgrade the performance of T&D construction are:

Design review Outage follow-ups Audits Construction upgrade projects

The objective is to start with a fault-resistant design, review outages to prevent repeat faults, audit to help ensure crews build to the design, and implement programs to correct past construction deficiencies. This chapter provides the framework for implementing these programs, and material provided in the rest of the report and in future work in this project series will help fill in details on specific construction practices. Initiation of many of these programs requires training for field engineers and personnel. It is best to do this training in the field, looking at actual circuits and construction practices. Once the programs are operational, the programs themselves act as a form of ongoing training.

4-6

Construction Issues

Data Analysis and Targeting Programs When designing and implementing programs to reduce faults, outage management databases can help identify problems leading to faults. They can help identify which circuits to target, which areas have the most problems with trees, which areas have the most problems with animals, and so on. Outage databases can also help judge the effectiveness of fault-reduction programs. While it would be nice to have a fault database that logged all faults, that is normally not possible, so we use the next best thing: the outage database. An outage database doesnt track all faultsonly those that lock out a protective device and cause a sustained interruption. Therefore, keep in mind that results from outage database analyses may not include as much information on temporary faults. Best use of an outage database requires implementing good outage cause codes and having crews report the cause codes correctly. Good outage cause codes should concisely provide information pertinent to the fault. A code for weather helps. For failed equipment, the equipment type can help, as can age, manufacturer, and material (these especially help for cable equipment faults and targeting cable-replacement programs). Outage codes can also attempt to separate the physical cause from any deficiencies in the construction that led to the fault. For example, a transformer may have failed during a lightning storm, but a missing lightning arrester was the defect that allowed the failure; this could be entered as a cause of lightning with a deficient construction code entered with a notation indicating arrester problem. More specifics help (as long as they are not too complicated). Taylor60 provides an example for tree outages; rather than labeling the cause trees, provide more specific options like: live tree in the right-of-way, dead trees and overgrown primary, trees far outside of the right-of-way, or trees on transmission lines. Outage databases can also be used to target repeat faults. Keep track of repeated outages on a protective device, especially those with similar cause codes. Some fault types have more repeatability than others. Animal faults can be highly repeatable at locations with poor clearances that are also high-traffic areas for animals. Tree vines can also cause repeated faults. The most important sections are not necessarily the locations with the most faults per mile. The number of customers on a circuit and the type of customers on a circuit are important considerations. For example, a suburban circuit with many high-tech commercial customers should warrant different treatment than a rural circuit with fewer, mostly residential and agricultural customers. How this is weighted depends on the utilitys philosophy. Fault location information can help when targeting programs or maintenance. Mapping fault locations can help on problem-circuit audits and similar investigations. Fault location is the primary factor that determines the disturbance severity to customers. Figure 4-4 shows several fault locations and how they impact a specific customer differently. A fault on the mains causes an interruption for the customer. If the fault is permanent, the customer has a long-duration interruption, but if the fault is temporary, the interruption is short as the protective device recloses successfully. A fault on a lateral tap causes a voltage sag unless fuse saving is used. With fuse saving, the fault on the tap causes a momentary interruption as the substation breaker
Taylor, L., The Illusion of Knowledge, Southeastern Electric Exchange Power Quality and Reliability Committee, Dallas, TX, 2003.
60

4-7

Construction Issues

or recloser tries to prevent the fuse from blowing (for discussions on fuse saving versus fuse blowing, see EPRI 100166561 or Short62 ).

subtransmission system

causes a voltage sag

distribution system

causes a momentary interruption or voltage sag (depending on use of fuse saving)

causes a voltage sag

Customer Location

causes a sustained interruption for a permanent fault or a momentary interruption for a temporary fault

Figure 4-4 Example Distribution System Showing Fault Locations and Their Impact on One Customer

Faults on adjacent feeders cause voltage sags, the duration of which depends on the clearing time of the protective device. The depth of the sag depends on how close the customer is to the fault and the available fault current. Faults on the transmission system cause sags to all customers off of nearby distribution substations. We can depict all of the possible fault locations by areas of exposure or areas of vulnerability as shown in Figure 4-5. Each exposure area defines the vulnerability for the specific customer. For sags, we have different areas of vulnerability based on the severity of the sag. An outline of the area that causes sags to below 50% is tighter than the area of vulnerability for sags to below 70%. We can use the area of vulnerability curves to help target maintenance and improvements for important sensitive customers.

61 62

Power Quality Improvement Methodology for Wires Companies, EPRI, Palo Alto, CA: 2003.1001665. Short, T. A., Electric Power Distribution Handbook, CRC Press, Boca Raton, FL, 2004.

4-8

Construction Issues

Sag Exposure Momentary Exposure

Customer Location

Sustained Interruption Exposure

Figure 4-5 One-Line Diagram Showing the Boundaries Where Faults Will Cause Sags, Momentaries, and Interruptions for the Customer Shown

Field Identification of Fault Sources Keeping in mind that most faults result from specific structural deficiencies, field identification of fault sources is a key part of many of the construction-improvement programs discussed here. Field personnel can be trained to spot pole structures where faults have occurred or might be likely. Several common structural deficiencies include:

Poor jumper clearances Old equipment (such as expulsion arresters) Bushings, arresters, or cable terminations unprotected against animals Ground leads or grounded guys near phase conductors Poor clearances with polymer arresters Damaged insulators Damaged covered wire Bad cutout placement 4-9

Construction Issues

Danger trees/branches present

Consider Figure 4-6. This is an obvious example of a location where a fault has occurred, probably repeated faults. The arrester and pole are severely blackened from arc burn products, and the transformer bushing has cracked and is missing a piece. This structure has several severe deficiencies:

The externally gapped arrester (which is possibly failed internally) provides a very short gap, easily flashed by birds or squirrels. And, it would only take a small lightning surge to flashover this structure. Neither the arrester nor the transformer bushing has animal protection. The guy wire attached near a phase conductor is uninsulated. The armless-type construction using post insulators provides little insulation on the pole. The equipment has no local fuse.

Overall, this structure is a disaster; it is highly susceptible to lightning, animals (even insects), small tree branches, and other debris. To fix this pole requires a complete overhaul: replace the arrester, replace the transformer bushing, add animal guards, add a guy insulator, add a local fuse, and reframe with crossarms or other materials to get more insulation and electrical clearances. While this example has many problems, any one of the problems by itself can be a source of faults, possibly repeated faults.

4-10

Construction Issues

Evidence of arcing on the arrester

Courtesy: Duke Power

Figure 4-6 Fault Arc Damage on a Gapped Arrester

Other examples can be much more subtle. Evidence of arcing from previous faults may be missing or hard to see from the ground (especially without binoculars). Inspections of structures to identify fault sources should be done from the ground (not while driving!). Implement training out in the field for best results; show examples of fault sources. Walk the line and use binocularsit is more effective than riding the line. Some fault sources are not obvious and require looking at a structure from different angles. When evaluating structures and possible fault causes, note the distinction between the cause of the fault and the deficiency. The cause of the fault may have been a squirrel, but the underlying source of the problem may have been poor electrical clearances (unprotected bushings, tight spacings, and so forth). We cannot get rid of the squirrels, but we can make structures more resistant to squirrel contacts. Figure 4-7 shows another example of a pole with some of the same problems: no animal guards, phase separations that are too tight, and there are no lightning arresters. While there are no obvious signs of previous arcs on this pole, it is a prime candidate for faults from lightning or animals.

4-11

Construction Issues

Source: EPRI 1001883

63

Figure 4-7 Transformer Pole With Minimal Phase-to-Phase Clearances and No Animal Guards

Figure 4-8 shows another example where arcing was evident. In this case, there is a rather short phase-to-phase spacing that is easily bridged by a squirrel or other animal. The arcing damage is obvious from the picture but still difficult to see from ground level. Corrective action could include raising the top dead-end insulator or routing the jumper under the crossarm rather than over it. In this example, the crossarm is very short, which limits options for maintaining high electrical clearances and insulation levels.

63

Distribution Wildlife and Pest Control, EPRI, Palo Alto, CA: 2001. 1001883.

4-12

Construction Issues

Tight phase-tophase spacings

Arcing evidence

Courtesy: Duke Power

Figure 4-8 Poor Clearances and Evidence of Arcing Across an Insulator

Damaged equipment should also be noted. Figure 4-9 shows an example of a blown arrester. Arresters normally fail to a short circuit, but if the arrester has a gap, the gap may be able to hold line voltage.

4-13

Construction Issues

Blown arrester

Courtesy: Duke Power

Figure 4-9 Blown Arrester

Design Review With Power Quality in Mind Fault-resistant designs are important for providing quality power. The first step in implementing fault-resistant designs is to start with good designs. Review all standard design drawings and specifications. Some questions to address in a design review are:

Are clearances appropriate to prevent animal contacts or contacts from other debris? Is the electrical insulation sufficient to withstand induced lightning surges? Are structures strong enough to withstand ice loads and moderate tree contacts?

During this research project, we will give more specific suggestions based on different threats to T&D circuits. In this report, for example, we cover the topics of animal protection and lightning protection. Future work will cover trees and other root causes of faults as well as information on specific structure types (riser poles, transformer poles, and so on). Fortunately, power quality design goals usually do not conflictbetter clearances for animalprotection also helps improve lightning protection. Keep in mind safety, workability, and cost as well as power quality impacts when deciding on design issues. In addition to reviewing existing designs, utilities may wish to review older designs to help identify past mistakes that need to be corrected. Many field site visits can also help utilities judge 4-14

Construction Issues

how existing designs are actually implemented and how past designs were implemented. This information will help utilities plan programs to upgrade deficient construction. Outage Follow-Ups On-site outage reviews can help identify weak points and reduce fault rates. Faults tend to repeat at the same locations and follow patterns. Consider animal faultsone particular pole, which happens to be a good travel path for squirrels, may have a transformer with no animal guards. The same location may have repeated faults. By identifying the location of fault and any structural deficiencies that contributed to the fault, the deficiencies can be corrected to prevent repeated faults in the future. The main goals of an outage-review program are:

Outage database improvementDid crews enter the fault code or outage code correctly based on available information? If not, feedback can correct the mistake, and lessons learned can help prevent future mistakes. TrainingField engineers and operating crews get more on-the-job training about specific design and construction deficiencies that lead to faults. SourceIdentify the most probable cause of the fault (tree limb, animal, lightning, and so on) Corrective actionConstruction deficiencies should be corrected to avoid repeat fault events. Corrective action is a prime goal of a fault review.

During an outage review, the main tasks of the reviewer are to review the data at hand and address the following questions:

What was the most likely location of the fault? What was the flashover path? What was the most likely cause of the fault? Was the outage code entered correctly? Was the faulted circuit adequately covered by protective devices? Was the fault on an unfused tap? Did protective devices operate as expected? What structural deficiencies contributed to the fault? Are the deficiencies likely to lead to additional faults? How could the deficiencies be corrected?

Based on addressing these issues, corrective action can be initiated if it is warranted. Corrective action could include rebuilds on one or more structures and could include tree/branch clearance. Corrective action could also include fixing unrepaired damage related to the fault.

4-15

Construction Issues

The first step is finding the most likely location of the fault that caused the outage. Permanent faults are relatively easy to pinpoint because the crew had to actually fix damage at one or more locations. Temporary faults are harder to pinpoint. If a fuse blows, the area narrows considerably. For areas with repeat fuse operations, careful patrols may identify areas where repeated faults occur. Still, if a tap fuse operates but is refused successfully, the cause may have been a squirrel across a bushing, a tree branch that fell onto then burned off of a line, or wind pushing two conductors together. It often takes a trained eye to determine the cause. Many outages are classified as unknown. For some outages, the cause will remain unknown. While the exact cause may remain unknown, information about the outage can help point the way to finding deficiencies. The time of day and the weather during the fault can help suggest a cause. Clear weather may suggest an animal fault, especially if it happens during the daytime. Stormy weather suggests lightning. An outage history for the protective device may reveal patterns. The probable cause can help the outage investigator look for deficiencies that could have led to the cause. Several options are available to decide which outages to review: all outages that impact over 500 customers (or some other number), all mainline outages, all outages on problem circuits, all outages on critical-customer circuits, outages on protective devices with excessive operations over a given time period (pick a threshold based on whether the device is a fuse, recloser, or circuit breaker), or a random portion of all outages. Utilities could also use combinations of these options in forming criteria for outage reviews. Responsibility for outage follow-ups will normally fall to a local field engineer. In addition to regular outage reviews, a utility may audit a sample of the outage reviews. A reliability auditor could visit the outage location with the local field engineer. The main point of such an audit is educate and reinforce consistent approaches to outage follow-ups. If at the time of the audit, the corrective action has been done, the auditor and the local field engineer can review the corrective action to determine if it was done properly. For example, check that animal guards were attached correctly and that sufficient clearances were maintained; other checks are discussed in the next section on construction audits. Construction Audits Never assume that crews will build according to specifications. Crews often make decisions based on misconceptions or because its the way weve always done it. Audits help reinforce practices to ensure that crews build to the specifications, and audits help educate crews on why things are done and how the designs minimize faults. Construction audits can apply to new construction but also to additions or changes to an existing structure, including work triggered by an outage follow-up. Options for a checklist for the auditor could include:

Animal guards installed correctly on all bushings and arresters Covered jumpers used

4-16

Construction Issues

Correct crossarm used Spacings acceptable per guidelines Guy wire insulated properly Riser bracket not grounded Proper connectors used Equipment tanks grounded properly Site tagged correctly Correct usage of current-limiting fuses Performed work specified No NESC construction grade violations Proper framing for the given structure angle Proper insulators used Span lengths do not exceed design limits Pole class not too small Proper conductor ties used Fuse barrels left hanging down in open cutout No grounds near the primary Arrester phase and ground leads per standards to minimize lead lengths Proper phase and neutral conductor sizes used

Of course, these need to be adjusted based on each utilitys construction design and historical construction practices and deficiencies. As the utility gains experience with the program, items can be added or taken from the list. With time, lessons learned during construction audits will spread through the company, construction will be more consistent, and the fault resistance of structures will increase. In addition to factors affecting a circuits power quality performance, other factors can be considered in a construction audit: NESC violations and other safety issues along with cost issues are the most common. For contract work, utilities can ensure that the contractors did the jobs that they took credit for. It is helpful to have a scoring system for audits. Tracking scores also reveals how well a district is doing over time. The auditor could deduct points based on a defined checklist. Scoring helps reinforce specific points, so penalize more severely those deficiencies more likely to be fault sources. Results are importantnot appearances. While it may be educational to point out sloppiness, that is not the main point. Bare jumper or ground leads can be very neatly applied on a structure 4-17

Construction Issues

but have insufficient clearances. In fact, sometimes the neatest looking path is not the path that gives the most electrical separation. Emphasize the most important issues that impact power quality and the fault resistance of a line: Are separations sufficient? Are animal guards applied correctly and securely? Are arresters mounted appropriately? Are arrester leads as short as possible? Who does the audits? Who is present at the audits? How often are audits done? What are the rewards or penalties (if any) based on audit performances? These are questions that utilities should work out themselves to ensure the most effective results within the constraints of a given organizational structure. Construction Upgrade Programs Outage follow-ups and construction audits along with reviews of outage data help pinpoint the weaknesses in existing construction. Utilities can shore up these weak areas by any one of several construction upgrade programs:

Animal-guard implementations Arrester replacements or new applications Tree clearance Danger tree programs Pole inspection and replacements Cable-replacement programs

Other options are possible, depending on past construction deficiencies. For example, if a utility has at some point built a number of lines with a shield wire (overhead neutral) that perform poorly, a program might target them for replacement or upgrades. A shield wire design can perform well if it has good insulation and grounding, but it can also perform poorly if not done rightthe grounding downlead can reduce insulation levels and make lightning contacts more likely and can increase the likelihood of animal contacts and contacts from other debris. Then, a shield-wire upgrade program could target these sections for replacement or for insulation upgrades to reduce animal and lightning faults. Equipment replacement programs also fall under the upgrade category. Many utilities have cable replacement programs. Program policies are done based on the number of failures (the most common approach), cable inspection, customer complaints, or cable testing. High-molecular weight polyethylene and older XLPE are the most likely candidates for replacement. Most commonly, utilities replace cable after two or three electrical failures within a given time period 64 (Tyner ). Because fault sources are most often at equipment poles, consider programs targeted specifically at equipment poles. For example, a transformer retrofit program could upgrade all transformer
Tyner, J. T., Getting to the Bottom of UG Practices, Transmission & Distribution World, vol. 50, no. 7, pp. 44 56, July 1998.
64

4-18

Construction Issues

locations to the utilitys design specifications. This could include replacement of old arresters with tank-mounted polymer metal-oxide arresters, animal guards on bushings and arresters, covered jumper leads, and elimination of grounded wires above the transformer tank (including grounded guy wires). It costs much less to upgrade construction when a crew is already set up at a pole. Consider implementing procedures and checklists for the crew to fix problems on the pole whenever they are set up for work on a pole. Options to consider in such a program are:

Add animal guards on unprotected bushings or surge arresters. Replace old surge arresters Add fiberglass guy insulators to uninsulated guys near the primary. Strip out unnecessary grounds above the neutral wire (especially those near primary conductors or jumpers). Replace any flashed or damaged equipment.

Problem-Circuit Audits Just as there are pole locations that can have repeated faults, faults can cluster on some circuits. These faults may be from consistently poor construction on a circuit, heavy tree exposure, or a few poor structures with repeated faults. The goals of a problem-circuit audit are:

SourceIdentify the deficiencies that are the most probably sources of faults. Corrective actionCorrect construction deficiencies to avoid repeat fault events.

Criteria for designating problem circuits can come from many options: circuits with critical customers that are having problems, circuits with excessive momentary interruptions, circuits with long-duration interruption problems (high SAIFI or SAIDI), or circuits with excessive customer complaints. The problem should also direct the focus of the audit. If the problem is momentary interruptions, review the section covered by the circuit breaker or recloser that is excessively operating. Pay special attention to animal and lightning susceptibility on the circuit. If the problem is long-duration interruptions, concentrate circuit patrols and reviews on the feeder backbone. Pay special attention to susceptibility to trees; a review of the outage records for the circuit should reveal even more information on where to look for problems. During the field review of the problem circuit, some deficiencies are obvious. Figure 4-10 shows a deteriorated pole top. Figure 4-11 shows a riser pole that has had multiple fault events. There are several paths that have very short clearances. The structure is highly susceptible to animal contacts, for faults both upstream and downstream of the fuse. The lightning protection is also less than desired; the phase and ground leads are longer than they could be. Both of these structures are obvious candidates for repair and upgrades. 4-19

Construction Issues

Figure 4-10 Deteriorated Pole Top

Figure 4-11 Riser-Pole Installation With Evidence of Multiple Arc Burns

Tree exposure is another major consideration in a problem-circuit audit (see Figure 4-12). The auditor must decide if the circuit warrants tree clearing. A danger-tree inspection is worthwhile to pinpoint trees and branches that are likely to fail. These include dead trees, leaning trees, or 4-20

Construction Issues

trees with major defects that threaten the line. Consider not just the immediate area around the conductors but also the overhang (Figure 4-13) and danger trees outside of the normal right-ofway but still within striking distance of the line. Decisions on tree maintenance should also include the time since the last tree maintenance on the circuit and the history of tree-caused interruptions on the circuit.

Figure 4-12 Trees Growing Into Lines

Figure 4-13 Lines With Significant Tree Overhang

Broken tree branches account for the majority of tree-caused interruptions. In one utility in the northeast United States, 63% were caused by broken branches compared to 11% from falling 4-21

Construction Issues

trees and only 2% from tree growth (Simpson65). Niagara Mohawk Corporation (NIMO) found that 86% of permanent tree-faults were outside of the right-of-way, and most were from major breakage (Finch66). Falling trees do the most damage; they often break conductors and even poles in some cases. Tree faults usually occur during storms, primarily from wind. Snow and ice additionally contribute to tree failures. The 2004 project will discuss tree faults and cutting practices in more detail. Some specific items for a repair checklist could include:

Animal guards missing on any surge arresters or bushings Transformers without local, external fuses Guy wires without proper insulation Pole ground wires run near the primary Old arresters Heavy tree coverage Danger trees Overhanging dead limbs Damaged insulators or bushings Deteriorated or damaged equipment Poor clearances Vines on equipment poles

Along with a problem-circuit construction audit, a utility should also review the protection schemes on a circuit and review other ways to reduce the impacts of faults. Are reclosers and other devices coordinated properly? Does the circuit have enough protective devices? Are there any unfused taps (even very short sections can cause problems)? For more information, see EPRI 67 1001665 .

Simpson, P., EUA's Dual Approach Reduces Tree-Caused Outages, Transmission & Distribution World, pp. 22 8, August 1997. 66 Finch, K., Understanding Tree Outages, EEI Vegetation Managers Meeting, Palm Springs, CA, May 1, 2001. 67 Power Quality Improvement Methodology for Wires Companies, EPRI, Palo Alto, CA: 2003.1001665.

65

4-22

Construction Issues

High-Tech Designs
Some utilities use robust, high-tech distribution circuits because of customer power quality needs. Constructions vary; options for more robust overhead constructions can include:

Shield wires Surge arresters Stronger structures Animal resistance measures Enlarged rights-of-way

A high-tech overhead circuit can have a much lower fault rate than traditional designs. If electrical clearances are large enough, and trees are cleared from the right-of-way, not many fault sources remain. A robust, high-tech distribution line with aggressive tree clearance (built like a subtransmission line) should have faults mainly from lightning and equipment failuresnot animals, not trees, not ice, and not wind. Also, a utility can prevent a substantial number of lightning-caused faults by adding a shield wire and/or surge arresters along with good construction practices. Fault rates for subtransmission lines can give us some idea of the fault rate achievable on a hightech overhead circuit. Traditional overhead distribution circuits average about 90 faults/100 68 miles/year. For 110- to 138-kV circuits, McGranaghan and Roettger cite an average fault rate of 7.2 faults/100 miles/year based on a survey that included 149,605 mile-years of survey data for that voltage range. Subtransmission lines have most of their faults from lightning and insulator contamination. High-tech overhead distribution lines could have similar rates, maybe not below 10 faults/100 miles/year, but 10 to 20 faults/100 miles/year might be achieved with robust designs. Because distribution lines have more equipment, equipment failures would likely compose a larger share of faults than those on subtransmission lines. Industry Practices Industry practices vary on the use of high-tech or high-reliability lines. In an informal survey of ten utilities, most did not have a formal high-reliability line design. The utilities that did were in the southeastern United States. Of the utilities that had a high-reliability design, usage was sporadic. Most uses are for industrial customers. And, most of the customers pay for some or all of the cost of the circuits. Most of the utilities that had high-reliability standards did not have performance data to compare with standard designs. Figure 4-14 shows one utilitys aggressive approach to a high-reliability design. This design uses 69-kV post insulators, a shield wire grounded at every pole, and line arresters on every third pole. The utility also attempts to have a 50-ft right-of-way (30-ft minimum) on both sides of the circuit. Long fiberglass guy insulators, large phase insulators, and fiberglass standoffs for the
McGranaghan, M. F., and Roettger, B., Benchmarking International Transmission System Fault Performance, PQA Conference, 2001.
68

4-23

Construction Issues

ground lead all help maintain high insulation. This design has a critical flashover voltage of at least 450 kV. As long as proper animal protection is maintained at equipment poles and the grounding is reasonable, this design should have a very low fault rate.

Figure 4-14 High-Reliability Design Example

Duke Power Case Study Duke Power has an outage-resistant design standard. Figure 4-15 shows the single-phase construction, and Figure 4-16 shows the double-circuit version. As of 1996, Duke had seven of these circuits, mainly feeding industrial parks or major customers. Normally, Duke offers this 4-24

Construction Issues

service to customers or industrial parks that are relatively close to the substation (within four miles). Their design uses several features for increased fault resistance:

1/0 ACSR shield wire Covered phase conductors (556.5 AAC) Surge arresters at 1000-ft (305-m) spacings Class 3 pole Every pole grounded; 25 or less preferred Ground wire routed away from phases to maximize insulation: uses fiberglass downlead standoffs on the double-circuit design Fiberglass guy strain insulators Arc-protective devices (L-nodes) on locations where the covered conductors are stripped for preventing conductor burndowns

Courtesy: Duke Power

Figure 4-15 Duke Power Outage-Resistant Design

4-25

Construction Issues

Courtesy: Duke Power

Figure 4-16 Duke Power Double-Circuit Outage-Resistant Design

Dukes outage-resistant design has performed well. Table 4-3 summarizes the performance and cost of the design. Note that the outage-resistant design is still a small sample, but it does show that significant performance gains are possible in a high-tech design.
Table 4-3 Performance and Cost Comparisons of Duke Powers Double-Circuit Outage-Resistant Design Standard Overhead Circuit Circuit breaker trips per year Cost per mile Note: 1996 data 1.6 $83,346 OutageResistant Design 0.1 $110,370 Multiplier

0.063 1.320

Underground Systems Underground systems are also high-tech options for circuits needing high performance. Underground circuits have fewer failures than traditional overhead T&D circuits. In addition to power quality performance, overhead and underground systems have many other differences that need to be considered for each application (see Table 4-4). The major advantage of overhead circuits is cost; an underground circuit typically costs anywhere from 1.5 to 5 times the

4-26

Construction Issues

equivalent overhead circuit. But the cost differences vary wildly, and it is often difficult to define equivalent systems in terms of performance.
Table 4-4 Overhead Versus Underground: Advantages of Each Overhead CostOverheads number one advantage. Significantly less cost, especially initial cost. Longer life30 to 50 years versus 20 to 40 for new underground works. ReliabilityShorter outage durations because of faster fault finding and faster repair. LoadingOverhead circuits can more readily withstand overloads. Underground Reliability and power qualitySignificantly fewer faults: fewer voltage sags and fewer short- and long-duration interruptions. AestheticsMuch less visual clutter. SafetyLess chance for public contact. O&MNotably lower maintenance costs (no tree trimming). Longer reachLess voltage drop because reactance is lower.

Underground circuits have fewer faults. Traditional overhead circuits typically fault about 90 times per 100 miles per year; underground circuits fail less than 10 or 20 times/100 miles/year. Because overhead circuits have more faults, they cause more voltage sags, more momentary interruptions, and more long-duration interruptions. Even accounting for the fact that most overhead faults are temporary, overhead circuits have more permanent faults that lead to longduration circuit interruptions. Cable circuits also have fewer multiple-phase faults than overhead circuits, which impacts power quality because multiple-phase faults cause more severe voltage sags in customer facilities. The main reliability disadvantage of underground circuits is that when they do fail, finding the failure is harder, and fixing the damage or replacing the equipment takes longer. This can partially be avoided by using loops capable of serving customers from two directions, by using conduits for faster replacement, and by using better fault-location techniques. Underground circuits are much less prone to the elements. A major hurricane may drain an overhead utilitys resources, crews are completely tied up, customer outages become very long, and cleanup costs are a major cost to utilities. Underground circuits are not totally immune from the elements. In heat storms, underground circuits are prone to rashes of failures. Underground circuits have less overload capability than overhead circuits; failures increase with operating temperature. In addition to less storm cleanup, underground circuits require less periodic maintenance. Underground circuits do not require tree trimming, easily the largest fraction of most distribution operations and maintenance budgets. The CEA69 estimated that underground system maintenance averaged 2% of system plant investment, whereas overhead systems averaged 3 to 4% or as much as twice that of overhead systems.

69

CEA 274 D 723, Underground Versus Overhead Distribution Systems, Canadian Electrical Association, 1992.

4-27

Construction Issues

Comparing the cost and performance of an underground line to a traditional overhead line or to a high-tech overhead line is difficult. Table 4-5 is a generic comparison based on the best industry data available, but note that many of the data have significant variation. Also, some of the data in Table 4-5 is based on limited industry reporting (especially for the high-tech overhead circuit but also for underground circuits). The performance level of high-tech overhead circuits depends on how robust the design is and how well the right-of-way is maintained. The performance of cable systems is also highly uncertain. Most cable-circuit failures are due to cable or splice failures; the industry consensus seems to be that modern cables and splices are more reliable than 1970s or 1980s vintage equipment, but no one knows for sure.
Table 4-5 Performance and Cost Comparisons Traditional Overhead Fault rate/100 circuit miles/year Long-duration interruption rate/100 circuit miles/year Average outage duration, hours (major events excluded) Average outage duration, hours (major events included) Cost, per unit, relative to traditional overhead References: Overhead fault rates: Figure 2 Cable failure rate compilations: Short70, Thue71 Overhead versus underground costs: CEA72 General overhead versus underground comparisons: Johnson73 90 50 1.3 2.3 1.0 High-Tech Overhead 20 10 1.3 2.3 1.4 Underground 15 15 2.5 2.5 25

On underground versus overhead versus high-tech overhead, each particular application must be evaluated individually. Costs of overhead and underground circuits vary significantly based on the terrain and obstacles and existing electrical infrastructure at a location under consideration. Performance needs also vary based on the customers served by the circuit.

Conductor Spacing and Span Lengths

Conductor spacing and span lengths play a critical role in determining how sensitive the system is to faults induced by wind, trees and the magnetic forces associated with short circuit currents. This section focuses on three main areas:
70 71

Short, T. A., Electric Power Distribution Handbook, CRC Press, Boca Raton, FL, 2004. Thue, W. A., Electrical Power Cable Engineering, Marcel Dekker, Inc., 1999. 72 CEA 274 D 723, Underground Versus Overhead Distribution Systems, Canadian Electrical Association, 1992. 73 Johnson, B. W., Out of Sight, Out of Mind? A Study on the Costs and Benefits of Undergrounding Overhead Power Lines, Edison Electric Institute, 2003. 11/02/03 DRAFT.

4-28

Construction Issues

1. National Electrical Safety Code Clearance Requirements and Span Lengths 2. Faults Induced by Short Circuit Forces 3. Wind and Tree Related Faults Overhead distribution system designs must address all of these areas affectively to provide reliable and safe operation. NESC Clearance Requirements A basic requirement of an overhead distribution system design is to provide a suitably robust system construction design and sufficient clearances of conductors to avoid faults and safety problems. This includes the spacing between phases, phase and neutral, neutral and ground, and any phase and ground. It also includes spacing between adjacent power supply circuits, communication circuits, and circuits on other supporting structures, as well as objects around and underneath the lines. The National Electrical Safety Code (NESC,74) has specific requirements for the minimum spacing and clearances of aerial conductors. The requirements of the NESC are intended to offer basic provisions for safeguarding of persons from hazards arising from the installation, operation, and maintenance of overhead electric supply and communication lines. While not directly focused on achieving a specific level of reliability and power quality, NESC requirements achieve this indirectly by establishing a certain level of resistance against some of the most common causes of line faults. However, it is important to recognize that the clearances and practices in the NESC are minimum requirements, and while they provide suitable levels of safety, they may not provide the desired level of reliability and power quality. The NESC was never intended for that purpose and its requirements may need to be substantially exceeded to achieve suitable results in many cases. The key areas of the NESC requirements that relate to conductor spacing and clearances for overhead distribution lines are detailed in Part II, Section 23, pages 69 through 152 of IEEE/ANSI Standard C2-2002. In this section, the following topics (or rules as they refer to them) are covered: a. Rule 230 General clearance issues b. Rule 231 Clearances of supporting structures c. Rule 232 Vertical clearances of conductors above ground, roadway, rail or water surfaces d. Rule 233 Clearances between wires, conductors and cables carried on different supporting structures e. Rule 234 Clearances of wires, conductors and cables from buildings, bridges, rail cars, swimming pools and other installations f. Rule 235 Clearance of wires, cables and conductors carried on the same supporting structures
74

IEEE C2-2002, National Electrical Safety Code.

4-29

Construction Issues

g. Rule 236 Climbing Space h. Rule 237 Working Space i. Rule 238- Vertical clearance between communications and supply facilities located on the same structure The most relevant spacing and clearance requirements as they relate to power quality and reliability impacts of trees, wind and fault current forces are in Rule 235 Clearances of wires, cables, and conductors carried on the same supporting structures. These clearances, to a great extent, determine the susceptibility of the circuit to the most common types of conductor faults relating to wind, trees and short circuit forces. In Rule 235, there is a certain minimum clearance required between each conductor based on operating voltage. In addition, there are increased clearance requirements based on the sag of the conductors. In the NESC method, the greater the line sag, the more the presumed amount of movement due to wind and other factors that may cause the conductors to establish momentary contact. In other words, the larger the sag, the greater the needed clearances. Also, NESC recognizes that smaller conductors tend to move more easily and need greater spacing for a given amount of sag so they have developed two tiers of clearance requirements one set for conductors less than AWG No. 2 and the other for conductors AWG No. 2 and larger. The NESC utilizes the following equation to determine the horizontal clearance requirements for conductors smaller than AWG No.2:

Clearance = 0.3 kV + 4.04 S 24

Where: Clearance is in inches kV is the RMS voltage between conductors S is the sag in inches

Eq. 4-1

Horizontal clearance requirements based on the above equation are shown in Table 4-6 for various amounts of conductor sag. These are the same clearances as specified in Table 235-2 of the NESC C2-2002, but they have been extended here to show 69 and 115 kV values (the NESC Table only goes to 46 kV).

4-30

Construction Issues

Table 4-6 Horizontal Clearances Between Line Conductors Smaller Than AWG No. 2 at Supports, Based on Sags (Adopted From NESC C2-2002 Table 235-2) Nominal Voltage Between Conductors (kV RMS) 2.4 4.16 12.47 13.2 13.8 14.4 24.94 34.5 46 69* 115* Sag in Inches 36 48 72 96 120 180 240 But Not Less than

Horizontal Clearance (inches) 14.7 15.3 17.7 18.0 18.1 18.3 21.5 24.4 27.8 34.7 48.5 20.5 21.1 23.5 23.8 23.9 24.1 27.3 30.2 33.6 40.5 54.3 28.7 29.3 31.7 32.0 32.1 32.3 35.5 38.4 41.8 48.7 62.5 35.0 35.6 38.0 38.3 38.4 38.6 41.8 44.7 48.1 55.0 68.8 40.3 40.9 43.3 43.6 43.7 43.9 47.1 50.0 53.4 60.3 74.1 51.2 51.8 54.2 54.5 54.6 54.8 58.0 60.9 64.3 71.2 85.0 60.1 60.7 63.1 63.4 63.5 63.7 66.9 69.8 73.2 80.1 93.9 12.0 12.0 13.5 13.8 14.0 14.3 18.5 22.4 26.9 NA NA

*Note: for voltages above 50 KV NESC requires use of the maximum voltage rather than nominal voltage in calculating the spacing requirements. Maximum voltage is typically 5% higher than nominal per ANSI C84.1-1995. In this chart the clearance based on the nominal voltage is shown for illustration purposes.

For conductors equal in size or larger than AWG No. 2, the following equation is utilized:
Clearance = 0.3 kV + 8 s 12
Eq. 4-2

Where: Clearance is in inches kV is the RMS voltage between conductors S is the sag in inches Horizontal clearance requirements based on Eq. 4-2 are shown in Table 4-7 below. These are the same clearances as specified in Table 235-3 of the NESC C2-2002. The clearance requirements for these larger conductors are somewhat less than the smaller ones. 4-31

Construction Issues Table 4-7 Horizontal Clearances Between Line Conductors AWG No. 2 or Larger at Supports, Based on Sags (Adopted From NESC C2-2002 Table 235-3) Sag in Inches Nominal Voltage Between Conductors (kV RMS) 36 48 72 96 120 180 240 But Not Less than

Horizontal Clearance (inches) 2.4 4.16 12.47 13.2 13.8 14.4 24.94 34.5 46 69* 115* 14.6 15.1 17.6 17.8 18.0 18.2 21.3 24.2 27.7
34.6 48.4

16.7 17.3 19.7 20.0 20.1 20.3 23.5 26.4 29.8

36.7 50.5

20.2 20.8 23.3 23.5 23.7 23.8 27.0 29.9 33.3

40.3 54.1

23.3 23.8 26.3 26.5 26.7 26.9 30.0 32.9 36.4

43.3 57.1

26.0 26.5 29.0 29.2 29.4 29.6 32.8 35.6 39.1

46.0 59.8

31.7 32.2 34.7 34.9 35.1 35.3 38.4 41.3 44.8

51.7 65.5

35.6 37.0 39.5 39.7 39.9 40.1 43.2 46.1 49.6

56.5 70.3

12.0 12.0 13.5 13.8 14.0 14.3 18.5 22.4 26.9 NA NA

*Note: for voltages above 50 KV NESC requires use of the maximum voltage rather than nominal voltage in calculating the spacing requirements. Maximum voltage is typically 5% higher than nominal per ANSI C84.1-1995. In this chart the clearance based on the nominal voltage is shown for illustration purposes.

In cases where sag is so minimal as to result in a very small clearance value, then the clearances cant be less than the minimum indicated amount in the preceding tables (see Table 4-6 and Table 4-7 rightmost column). For voltages above 50 kV, the NESC does not specify a minimum no sag horizontal clearance requirement so the rightmost column is not applicable for voltages above 50 kV. Another clarification is that the clearances should be based on the maximum rather than nominal voltage level for voltage levels above 50 kV. Voltages below 50 kV use the nominal voltage. In Table 4-6 and Table 4-7 above, the nominal voltages are used at 69 and 115 kV for illustration purposes only and for a real world application the maximum voltages would need to be used resulting in slightly larger clearance requirements than shown for those two rows. Normally, the maximum voltage would be about 5% higher than the nominal voltage per the ranges specified in ANSI C84.1-1995. For voltages exceeding 470 kV, the NESC requires that the horizontal clearances be determined by an alternative equation (see NESC C2-2002 for details). There are also vertical clearance requirements. Vertical clearance requirements for vertically arranged conductors are a bit less than those for horizontally arranged conductors. A reason for 4-32

Construction Issues

this is that wind and fault induced motion on sagging vertically arranged conductors will not cause them to contact each other the way horizontally sagging conductors can be affected. On the other hand, such vertically arranged conductors can sag into each other if thermal or mechanical loading is considerably unbalanced (between phases) or if one conductor breaks or due to unbalanced conductor tensions (due to conductor breakage, mechanical issues, etc.).
Table 4-8 Vertical Clearances Between Conductors at Supports Open Supply Conductors Conductor Type (usually at lower level) Over 8.7 kV to 50 kV 0 to 8.7 kV (inches of spacing)
Open conductors less than 750 V (such as secondary) or effectively grounded neutrals Open conductors 750 V to 8.7 kV Open Conductors 8.7 to 22 kV: 1. Worked on energized, but with adjacent circuits energized 2. Worked on energized, but with adjacent circuits deenergized Open Conductors 22 kV to 50 kV 16 plus 0.4 per kV over 8.7 kV 40 plus 0.4 per kV over 8.7 kV

Same Circuit (utility) (inches of spacing)

Different Circuit (Utility) (inches of spacing)

16

16 plus 0.4 per kV over 8.7 kV

40 plus 0.4 per kV over 8.7 kV

16

16 plus 0.4 per kV over 8.7 kV

40 plus 0.4 per kV over 8.7 kV

16 plus 0.4 per kV over 8.7 kV

16 plus 0.4 per kV over 8.7 kV

16 plus 0.4 per kV over 8.7 kV

16 plus 0.4 per kV over 8.7 kV

The clearance requirements shown above for vertically arranged conductors are at the supports. NESC recognizes that the clearance might be less at midspan and requires that the sag on all conductors shall be such that clearances at any point on the span shall be no less than 75% of the clearances specified at the supports. Note that there are special clearance requirements in Table 4-8 associated with the ability to allow live line work without de-energizing an adjacent circuit. The NESC Section 23 also specifies the clearances of conductors above roadways, railroads, buildings, pools, waterways, walkways, etc. There are several tables and rules applying to different types of structures and objects. Table 4-9 shows some selected excerpts under Rule 232 of the clearance requirements in the NESC for vertical clearances above the ground, roadways, rail tracks, and water. The clearance used in these tables is the worst of the following: the clearance that occurs at a conductor temperature of 50 degrees C (120 F) with no wind, or at the 4-33

Construction Issues

maximum conductor operating temperature (if a higher temperature) with no wind, or at 0 degrees C (32 F) with the maximum design ice loading condition and no wind.
Table 4-9 Vertical Clearance Requirements Above Ground, Roadways, Rail Track and Water Supply Conductors Under 750 volts (Feet) Open Supply Conductors 750 volts to 22 kV Line-to-Ground (Feet) 26.5 18.5 18.5 18.5 14.5 17.0

Nature of Surface Beneath Wires, Conductors or Cables

1. Track rails of railways (except overhead electrified railways) 2. Roads, streets and other areas subject to truck traffic 3. Driveways, parking lots, and alleys 4. Other land traversed by vehicles (cultivated grazing forests, etc.) 5. Spaces and ways subject to pedestrians 6. Water areas not suitable for sailboats 7. Water areas suitable for Sailboats: Less than 20 acres Over 20 to 200 acres Over 200 to 2000 acres Over 2000 acres

24.5 16.5 16.5 16.5 12.4 15.0

18.5 26.5 32.5 38.5

20.5 28.5 34.5 40.5

8. Established boat ramps and associated rigging areas

Clearance above ground shall be 5 ft greater than in 7 above for the type of water areas served by the launching site

Where wires conductors or cables run along and within limits of highways or other road rights-of-way but do not overhang the roadway 9. Roads, streets, or alleys 10. Roads in rural districts where it is unlikely that vehicles will be crossing under the line 16.5 16.5 18.5 18.5

Covered Conductors and Neutrals The required clearances for covered conductors (such as tree wire) are considered to be the same as bare (open) conductors for all clearance requirements except that spacing may be reduced below requirements for open conductors when the conductors are owned, operated or maintained 4-34

Construction Issues

by the same party and when the conductor covering provides sufficient dielectric strength to limit the likelihood of a short circuit in case of momentary contact with other covered conductors or the grounded conductor. Neutral conductors of circuits with 0 to 22 kV line-to-ground voltage that are effectively grounded are treated with the same clearances as guy wires and messenger wires. All other neutrals of supply circuits shall have the same clearances as phase conductors. Illustration of Clearances The application of the clearances in Table 4-6 through Table 4-9 is illustrated graphically in the following diagrams. Figure 4-17 shows the application of horizontal clearances at the supporting crossarm as well as the vertical clearance above a roadway. The vertical clearances are measured at the point of minimum clearance under the worst-case sag conditions discussed earlier. Figure 4-18 shows the application of vertical clearances between open conductors of the same circuit.

Horizontal Clearance at Supports

Sag

Effectively Grounded Neutral Clearance to Road Surface

Phase Conductor Clearance to Road Surface

Road

Figure 4-17 Application of Horizontal Clearances at Supports Between Conductors and Vertical Clearance Between a Road and Conductors

4-35

Construction Issues
Vertical Clearance at Supports Clearance Should be No Less Than 75% of Clearance at Supports

Vertical Clearance between Neutral and Ground

Vertical Clearance between Phase and Ground Surface

Figure 4-18 Vertical Clearances Between Conductors

A complication in determining the required clearances is that not all conductors are displaced just horizontally or just vertically. For example, candle stick or armless arrangements like that shown in Figure 4-19 have both a horizontal and vertical component of spacing. In such cases, the required spacing is determined by applying both the vertical and horizontal clearance requirements of the NESC to each axis respectively.
Horizontal Clearance Horizontal Clearance

Vertical Clearance

Figure 4-19 Geometries Involving Both Horizontal and Vertical Clearances

4-36

Construction Issues

Climbing Spaces and Work Areas The NESC addresses climbing spaces and working areas on poles and between wires. These clearances are intended to provide sufficient space between conductors, equipment, and/or circuits such that a line crew can safely perform maintenance or restoration work and can safely access the needed work areas without causing a fault or exposing themselves to undo danger. For example, the horizontal clearance between conductors bounding a climbing space on a 13.2-kV circuit between phases should be no less than 30 inches according to Rule 236. On a 25-kV class circuit this becomes 36 inches and on a 34.5-kV circuit it is 40 inches. Full details on climbing spaces and work areas can be found in Rules 236 and 237 in Part 2, Section 23 of the NESC. Figure 4-20 shows an example of a climbing space between several feeder circuits on a common pole. If the climbing and working spaces are sufficient, it is possible for the crew to work on one circuit, while the other two remain energized. This not only facilitates faster restoration for the failed circuit, but also avoids an interruption on the other circuits that would otherwise be needed to allow crews to restore service on the failed circuit. Furthermore, proper working spaces reduce the chance crews will accidentally cause a fault on an energized circuit during restoration improving safety for crews as well as power quality and reliability for customers.

Circuit 3

Circuit 1 Horizontal Climbing Space Clearance

Circuit 2

Figure 4-20 Example of a Climbing Space That Has Been Provided in the Layout of the Pole to Allow Safe Crew Access

Span Lengths Span lengths for medium and high voltage supply conductors of distribution lines are not directly addressed in the NESC using a specific table of span length requirements. Instead, they are addressed indirectly through requirements for the maximum allowable conductor tension as a percentage of its rated breaking strength. This approach works fine because the physics of supporting a wire on poles automatically establishes a needed span length if a target tension level 4-37

Construction Issues

must not be exceeded for a given wire size. The support poles will need to be spaced at specific intervals for each type of wire and external load conditions so for structures carrying multiple utilities and circuits, the weakest type of conductor will establish the span length requirement. The amount of tension for a given span is a function of the wire size and type, ancillary wire hardware weight (such as spacers, clamps, etc), and external loading due to wind and ice conditions. These factors, along with the grades of line construction, mechanical strength requirements and wind/ice loading requirements determined the appropriate span length and are discussed in detail in Sections 24, 25 and 26 of the NESC. The allowed tensions include factors for ice, snow, and wind loading. The NESC has specific maps and several tables of loading factors for wind and ice in various regions of the country. For example, in heavy ice loading areas, the design is based upon inch of radial ice collection on the conductors. In medium and light icing areas 0.25 and 0 inches of ice collection is the minimum design requirement respectively (see Rule 250). The wind speed design criteria ranges from 85 mph in some western areas to 150 mph in certain coastal regions of the southeast. There are several different grades of construction depending on the type of circuit to be carried (communications, supply conductors, etc.). The basic design load requirement (which includes the internal load which is the weight of the conductor and parts plus external loads such as ice and wind) should not exceed 60% of the rated mechanically loaded breaking strength per NESC Rule 250B. Furthermore the tension, without external load should not initially exceed 35% of rated breaking strength, and after settling in the unloaded tension should not exceed 25%. There are special cases where overload factors are applied to the design loading requirements. These factors range from under 1.0 to 4.0 depending on the application and are meant for certain part of the line and at certain locations with higher stress or criticality. The specific span lengths used by utilities are as much a function of economics as they are NESC requirements and other technical factors. The locations of customers also play a key role in dictating pole placement and hence span length. Generally in rural areas spans for distribution primary construction may be anywhere from 40-200+ meters, whereas in urban/suburban areas spans are typically only 30-60 meters. The numerous equipment and customer locations in urban areas dictate shorter spans. Also, those areas are likely to have more CATV and telephone services that may require tighter spans (note: the NESC also has requirements that address CATV and telephone cable spans). If economic consideration was the only factor of interest and customers were far apart, then very large span lengths pushing the conductor tensions, sags and clearances to the NESC limits would be the most desirable approach for distribution circuits. However, since power quality and reliability are also a concern, the most economic design may not be the best from an overall performance perspective. Smaller spans allow for a stronger distribution line that is less likely to be damaged by wind and ice (because each pole takes a smaller portion of the conductor wind/ice load and conductor tensions can be less). Furthermore, tighter spacing can reduce conductor galloping due to wind and ice and also conductor slapping due to forces associated with fault currents. If the span length is reduced from the maximum possible, faults can likely be reduced to a certain extent. However, this only occurs to a certain lower limit of span length. As span length is further reduced, there are a significantly increased number of poles that result in more opportunities for automobile accidents into poles or pole damage. There are also be more 4-38

Construction Issues

opportunities for animal or bird faults at pole tops and insulator flashovers as pole-tops. Obviously, there is an optimum span length from a power quality and reliability perspective. For rural systems, this length is likely to be shorter than the optimum economic length, but for suburban systems with short spans it might actually be longer. There would be an interesting area of focus for a future study to address the optimum span length for reliability based upon all the causes of faults. In considering the NESC clearance requirements and span distances that can be achieved, there are creative solutions that can help reduce faults while still maintaining long spans for economic purposes. For example, use of a longer crossarm to increase phase conductor clearances can decrease conductor slapping or wind related contact faults. Use of an increased spacing between phase and neutral can reduce contact between phase and neutral lines while allowing span distances to be maintained. For vertically oriented lines, increased vertical separation beyond the NESC minimum requirements can reduce the susceptibility to icing faults. Neutrals built under lines can be offset using a small crossarm or bracket so the horizontally oriented phase conductors above dont sag into them if burdened with ice. Certain types of unconventional conductor spacing patterns such as wide triangles (similar to candlestick layouts only larger) may help reduce faults. It is important to recognize that the NESC limits are the minimum that have been established mainly for safety purposes and not for economic or reliability reasons. The line designer has the flexibility to exceed the NESC clearance and design requirements and use innovative layouts when these can lead to better economics or reliability in the line construction.

Conductor Slapping Due to Short Circuits

Conductor slapping due to short circuits is a phenomenon caused by the magnetic forces associated with short circuit currents flowing in the line. Depending on where a fault is located on the system, short circuit currents can be 10-100 times larger than typical load currents and the onset of a fault will result in a significant magnetic force between the phase conductors. This force can cause substantial conductor movement. If the conductors are spaced too closely (poor clearances) and/or if there is too much play (easily achieved due to high sag) a short circuit at one location may trigger enough motion in a conductor upstream to cause a subsequent momentary fault at another location. The lines that are most easily affected are those with a combination of lighter weight conductors, tight conductor spacing, above average sag, long spans and relatively high fault levels. Any span where these characteristics occur together could be particularly susceptible to short circuit induced conductor slapping. Conductor slapping due to short-circuit current forces is not just an obscure problem, but instead one that is widely encountered at most utilities. Because of its transient nature and occurrence at typically unknown locations as well as lack of visual evidence following the event, many incidents simply go undetected and it is not really known in the industry what percent of faults also lead to conductor slapping faults. It can be difficult to identify that slapping has occurred even when it is known. For example, Allegheny Power investigated slapping on a 46-kV line using fault recorders to determine the location of the conductor slapping faults. Even with this information, visual inspection at the possible fault locations revealed only minor pitting/melting

4-39

Construction Issues

of the aluminum, and it was very difficult to recognize that slapping had occurred75. Had the slapping not been detected with fault recorders, these marks probably would have been discounted as being from earlier flashovers unrelated to conductor slapping. Conductor slapping is more than an inconvenience for the utility company. It can cause unnecessary operation of protection devices, unneeded lockout of circuit breakers or reclosers, wasted resources in tracking down perceived equipment coordination problems, and additional equipment deterioration (substation transformer subjected to more fault events, circuit breaker operations increased, conductors are unnecessarily pitted/partially melted). Overall, it degrades reliability and power quality and results in increased maintenance cost for the distribution system. The conductor-slapping phenomenon is illustrated in Figure 4-21. In this example a phase-tophase flashover between phases B and C at a pole top results in a strong magnetic force (initially repelling the two conductors). During the initial fault, the forces can be thought of as an impulsive force on the conductors swinging them outwards from each other. The magnetic force will usually be terminated well before the conductors reach the outward limit of their swing because the fault may last only 4-6 cycles with an instantaneous circuit breaker operation, and the crest of the swing outwards occurs much later. Once the crest of the outward swing is reached, then the conductors, each like a pendulum, will swing back in approaching each other and potentially faulting if there is insufficient clearance. Analysis shows that the line-to-line type of fault is normally the mode of fault that causes the worst forces and highest likelihood of a conductor-slapping event.
Phase B-C Initial Flashover. Fault at pole top
Conductors initially forced apart by magnetic forces, then swing back together and flashover.

ib ic

Phase A

Cross Section Showing Initial Forces on Conductors

Phase A Phase B Phase C

Initial Force

Initial Force

Figure 4-21 Illustration of Conductor Slapping Fault Due to Magnetic Forces of an Earlier Flashover

75

Frank, F. and Reese, J., "Arrestering Lightning Problems," Transmission & Distribution World, April 2001.

4-40

Construction Issues

Magnetic Forces We can calculate the short circuit forces on conductors from basic physics. The magnetic force between two parallel conductors is determined by the following equation adapted from the 76 Aluminum Conductor Electrical Handbook :

F =M
Where:

5.4 I 2 d 10 7

Eq. 4-3

M = Short circuit force multiplier (based on dc offset of fault current. Use 2 for a symmetrical fault and 8 for a fully offset fault) F = Pounds per linear foot of conductor I = Line-to-line short-circuit current in each conductor in amps d = Spacing between centerlines of conductors in inches A line-to-line fault with tightly spaced crossarm or candlestick construction is considered the worst type of fault from the point of view of magnetic forces on conductors. This is because a three phase fault results in a less intense magnetic field between any two conductors (due to field cancellation effects) and a single phase to ground fault results in interactions between the neutral and phase, which are spaced farther apart and usually vertically oriented resulting in much less force and conductor displacement. The one exception where the line-to-ground fault can be worse than a line-to-line fault is the case where the neutral is also on the crossarm in close proximity to the faulted phase, and the fault is at a location where the zero-sequence impedance is low (such as near the substation and with a delta to grounded/wye substation transformer). Plugging values of line-to-line fault currents ranging from 500 to 20,000 amperes into Eq. 4-3 above and assuming a typical M value of 4 (this is an average offset for illustration purposes), we can generate a table of results (see Table 4-10). We can see that the forces range from less than a tenth of a pound per foot at wide spacing and low currents to over 50 pounds per foot at close spacing and very high fault currents. As an example, a conductor with a spacing of 24 inches, and 5000 amperes of symmetrical fault level, would experience about 1.69 pounds of force per foot of conductor. On a 200-foot span, this is 338 pounds of force pushing the two conductors apart.

76

Aluminum Association, Ampacities for Aluminum & ACSR Overhead Electrical Conductors, 1986.

4-41

Construction Issues

Table 4-10 Force Per Linear Foot of Conductor for a Line-to-Line Fault on a Feeder
Phase to Phase Horizontal Conductor Spacing (inches) Line to Line Fault Level (Amperes) 12 24 36 48 60 72

Pounds of Force Per Linear Foot of Conductor 500 1000 1500 2000 3000 4000 5000 7500 10000 15000 20000 0.034 0.14 0.30 0.54 1.22 2.16 3.38 7.59 13.50 30.38 54.00 0.017 0.07 0.15 0.27 0.61 1.08 1.69 3.80 6.75 15.19 27.00 0.011 0.05 0.10 0.18 0.41 0.72 1.13 2.53 4.50 10.13 18.00 0.008 0.03 0.08 0.14 0.30 0.54 0.84 1.90 3.38 7.59 13.50 0.007 0.03 0.06 0.11 0.24 0.43 0.68 1.52 2.70 6.08 10.80 0.006 0.02 0.05 0.09 0.20 0.36 0.56 1.27 2.25 5.06 9.00

The characteristics that contribute to the likelihood of conductor slapping include:

High fault currents (faults closer to the substation) Maximum dc offset (faults at locations with high X/R ratio and at the point of wave causing highest offset) Long spans and close conductor spacing High levels of sag Smaller, lighter conductors Conductors in the same horizontal plane Increased duration of fault current

Dominion Resources recently published an analysis of the dynamics behind conductor slapping 77 behavior and the susceptibility of various construction types to the slapping phenomenon . They concluded that armless construction (such as candlestick arrangements) with their relatively tight spacing would be one of the more vulnerable types of construction. Table 4-11 shows the results of three different construction types that they analyzed for fault forces.

Ward, D. J., "Overhead distribution conductor motion due to short-circuit forces," IEEE Transactions on Power Delivery, vol. 18, no. 4, pp. 1534-1538, 2003.

77

4-42

Construction Issues Table 4-11 Short Circuit Forces for Typical Constructions Distribution Construction Type Armless 2.4 m (8ft) crossarm 3.0 m (10 ft) crossarm Source: Ward
78

Closest Phase Spacing 0.8m (32 inches) 1.1m (44 inches) 1.4 m (56 inches)

Per Unit Magnetic Force Compared to Standard 44 Inch Spacing 1.375 1.00 0.786

To determine how much the conductors move and whether or not a fault will result is much more complicated than simply calculating the steady state magnetic forces on the conductor. The fault event is a transient event, imparting what amounts to an impulse force to the conductor causing it to swing outwards to a crest point and then return in a harmonically damped motion. This motion is best compared to the swinging oscillations of a pendulum, and we can use the same dynamic equations that describe pendulum motion to evaluate the total swing of the conductors and whether or not a fault will occur. The first step in doing this is to simplify the conductor into a single lumped mass with the appropriate pendulum length. A pendulum with mass equal to 1 span of conductor and with an arm length equal to 2/3 of the conductor sag is a good representation of a swaying conductor for dynamic modeling purposes (see Figure 4-22)79.

Conductor Sag

2/3 Sag

Mass

Figure 4-22 A Sagging Conductor Can Be Represented by a Mass Equal to the Weight of One Span With an Arm Length Equal to 2/3 of the Total Sag of the Conductor

Based on harmonic oscillation concepts that we can apply to our representative pendulum we can calculate the range of motion (the amount of swing) of each conductor and the separation between them. The swing of both conductors plotted as a function of time is qualitatively illustrated in Figure 4-23. Here we can see that at fault initiation, the conductors begin moving apart accelerating as the fault continues. Upon fault clearing, the conductors inertia allows the
Ward, D. J., "Overhead distribution conductor motion due to short-circuit forces," IEEE Transactions on Power Delivery, vol. 18, no. 4, pp. 1534-1538, 2003. 79 Ward, D. J., "Overhead distribution conductor motion due to short-circuit forces," IEEE Transactions on Power Delivery, vol. 18, no. 4, pp. 1534-1538, 2003.
78

4-43

Construction Issues

separation to increase until a crest value of separation is reached. At this stage the conductors start falling back towards each other eventually reaching a point of minimum separation (or contact). If the minimum separation is small enough, then a flashover could occur. There will be some damping as shown in the illustration with each subsequent oscillation. The damping shown has been exaggerated for clarity.
Conductor Displacement Axis
Phase B Normal Position Point of closest approach where slapping fault may occur if air breaks down or contact occurs Phase C Normal Position Phase C Position First Crest - conductors have move to maximum separation distance Conductor motion is underdamped and takes many cycles of movement to dissipate

Phase B Position

Fault Initiation

Fault Clearing

Time Axis (seconds)

Figure 4-23 Theoretical Illustration of the Conductor Displacement Versus Time

From basic harmonic motion theory, we can calculate the period of a pendulum if we know the distance (l) of the arm and the force of gravity:

= 2
Where:

l g

Eq. 4-4

= the period of the oscillation in seconds

l=the length of pendulum in meters (that is 2/3 of the total sag of the conductor in meters or feet) g = gravitational constant (which is 9.8 m/s2if meters are used for sag or 32.2 ft/s2if feet are used for sag) From this formula we calculate the conductor oscillation period for various amounts of sag (see Table 4-12). Notice that the conductor weight and span length do not directly factor into the 4-44

Construction Issues

period of oscillation. Rather it is dependent entirely on the acceleration due to gravity (a constant) and the pendulum length.
Table 4-12 Period of Conductor Oscillation for Various Sags (Assuming no Damping) Total Conductor Sag (feet) 2 3 4 5 7 9 Period of Oscillation (seconds) 1.28 1.57 1.81 2.02 2.39 2.71

The periods of oscillation are long compared to the typical high current fault duration on a feeder. Most faults we are concerned about will result in instantaneous breaker operations (well under 10 cycles duration), and only some may be time delayed lasting up to 60 cycles perhaps. Few of concern would ever be more than 60 cycles. Given the period of oscillation, if there is an instantaneous trip, the fault will be cleared well before the conductor reaches its first crest (maximum separation). In fact there is more than enough time in these slow period oscillations for the high speed reclosing without intentional dead time delay to clear the fault and re-energize the line in time to allow a conductor slapping fault. It is a simple matter to determine the first swing conductor displacement in the X and Y coordinates by using basic physics and straightforward trigonometry. We can use the basic laws 2 of conservation of energy and Newtons laws of motion (F=MA and X=1/2AT ) to figure this out as an approximation. If we know the force input to the conductor, we can then calculate its acceleration. The force input is:

J = F dt
Where: J= impulse in pound-seconds F = force in pounds t = time in seconds

Eq. 4-5

From the force input, we can now calculate the acceleration of the conductor using F=MA (we know the mass of the conductor and the force applied). From F=MA, acceleration and speed are determined, and we can now calculate the kinetic energy of the conductor at the bottom of the 4-45

Construction Issues

swing angle, which is equal to MV2. Due to conservation of energy, the conductors kinetic energy must be transformed completely into potential energy (MGH) at the first crest of the swing. If we set MGH = MV2, we can find the height of the initial swing (H) as shown in Figure 4-24. Once we know the height (H), we can assume the conductor, acting a pendulum, will swing the same amount to the other side based on conservation of energy. (In fact, it wont swing quite as far due to losses but these can be ignored.) From this point, it is a simple matter of trigonometry to calculate the X and Y axis displacements and clearance between the conductors during the second crest.
Conductor Pivot Point Conductor Pivot Point

2nd Crest (minimum separation)

E= MGH
First Crest (maximum separation)

H
E= MV2 Initial Position
Phase B Initial Position Phase C

First Crest (maximum separation)

Figure 4-24 Shows the Method of Conservation of Energy to Calculate the Height H of the Initial Swing

A number of researchers have performed analysis of the conductor slapping phenomenon and several technical resources are now available. One resource is the EPRI Transmission Line Reference Guide: 115 138 kV Compact Line Design80. It has extensive discussion on conductor movements due to wind and magnetic forces for compact transmission line designsand even though it is transmission focused, much of the theory can also be applied at subtransmission and distribution levels of the system. Another possible source of information is Duke Power. Duke Power has developed an in-house software package called MISFAULT designed to calculate 81 fault forces and predict conductor swings . It has been used successfully for several studies. A recent IEEE paper written by Dan Ward of Dominion Resources is one of the best detailed papers on this topic available and is focused entirely on distribution applications82. Dominion Resources studied this problem extensively on many of their circuit designs and developed an inhouse software package for studying fault forces and conductor movements. They performed several parametric studies of the effects of fault currents on various distribution configurations to identify optimum circuit breaker/recloser settings and line design practices to minimize conductor slapping problems.

EPRI, Transmission Line Reference Book: 115 - 138 kV Compact Line Design, Second ed, Electric Power Research Institute, Palo Alto, California, 1978. 81 Craven, K. and Xu, X.-B., "Misfault Puts the Reins on Galloping Conductors," PQA Conference, May 2004. 82 Ward, D. J., "Overhead distribution conductor motion due to short-circuit forces," IEEE Transactions on Power Delivery, vol. 18, no. 4, pp. 1534-1538, 2003.

80

4-46

Construction Issues

Possible Conductor Slapping Solutions The work that has been done in the industry by EPRI and several utilities is providing a better understanding of this issue. Calculation methodologies based on computer software now exist to analyze this problem and determine the appropriate line design criteria to minimize slapping faults. It is recommended that utilities always include in any new distribution line designs the proper provisions in the design to limit conductor slapping. This could mean screening the basic design characteristics to look for span situations that are susceptible to conductor slapping. Use of technologies such as covered conductor can also solve the problem. By staying away from long spans, tight conductor spacing, and excessive sag, utilities can reduce this problem. Duke Power found that an easy but effective way of reducing the chance of conductor slapping faults is to always mount the middle-phase pin insulator on the pole rather than on the crossarm. In addition to gaining more horizontal separation, the additional vertical separation helps separate the conductor swinging motions. By changing the force vectors to include a vertical as well as a horizontal component, the force pushing the conductors apart is reduced. Ward83 provides some excellent graphs for common distribution feeder designs which were analyzed for slapping faults. While these results apply mainly to one companys designs, they are generic enough to be useful to many others. These results show the critical fault levels and clearing times where conductor slapping could become an issue for various conductor spacings and span lengths. The results suggest that most problems occur with time-delayed faults and that instantaneous trips would mitigate many problems. In cases where conductor slapping is an identified problem on already existing circuits, there are solutions available that utilities can consider. These include the installation of spacers at midspan to reduce conductor excursions on the impacted spans. In addition, shortening the duration of faults by using faster tripping times for circuit breakers and reclosers is another technique to reduce the total displacement of the conductors during the fault. Identifying that there is a problem is a key part of the process, and utilities can use monitoring resources (like digital relays) to look for signs of this problem. A telltale sign is when line-to-line faults occur that are followed within 1-2 seconds by a fault of larger magnitude. This would be an indication that a downstream fault has triggered a new fault closer to the substation probably as a result of slapping conductors.

Wind and Trees

Tree Faults With a tree branch across conductors, the amount of time it takes to go from a high-impedance fault to low impedance can range from a few seconds up to many days or even never in some cases. In some cases the limb will eventually fall away or move away from the line never causing a low-impedance fault. Usually a limb must bridge two conductors across a short distance (less than 4-6 feet) for a fault to be established. A growing limb touching a single conductor usually
Ward, D. J., "Overhead distribution conductor motion due to short-circuit forces," IEEE Transactions on Power Delivery, vol. 18, no. 4, pp. 1534-1538, 2003.
83

4-47

Construction Issues

has too much impedance in the current path back through the tree trunk and wont establish a sufficient voltage gradient to develop a low-impedance fault. Tests by ECI have shown that as long as the voltage gradient along the wood is less than about 2 kV/ft that it is unlikely that a low-impedance fault will develop when a branch bridges across 84 two conductors (see Goodfellow ). With the spacing found on typical 12.47- and 13.2-kV lines, such as candlestick construction or 8- to 10-foot crossarm construction, the voltage gradient would be large enough for a fault to develop. At 15 kV or less, bridging distance longer than about 6 feet should bring the gradient down to a level where a fault wont occur. This has some ramifications for line designs. It suggests that the use of a wider than normal crossarm configuration such as 12 feet could allow conductors to be spaced in a manner that would greatly reduce tree faults at the 15-kV class level. It would not be practical at the 25- and 35-kV class level because huge crossarms of 25 feet and over would be required. It is noteworthy that 4.16and 4.8-kV distribution designs see fewer tree faults of this nature. It is clear that the already existing spacing levels used in typical 5-kV class construction are effective at limiting bridging faults since the voltage gradient is typically well under 2 kV/ft in those situations. Covered conductor also solves these faults and is discussed later. There is also the question of how vertically oriented bare conductors perform in the tree branch bridging situation compared to horizontally oriented bare conductors. With vertical distribution designs, as long as the weight of the tree branch is not sufficient to force the upper conductor to sag into the lower conductor, it seems very plausible that this type of construction might perform better than horizontal construction for most types of bridging branch conditions. On the other hand, other types of tree faults such as heavy limbs that force conductors together (discussed later) may actually be worse with typical vertical conductor spacing. Since those types of faults appear to be more prevalent, vertical construction may be worse than horizontal from an overall tree fault perspective. Other than comparisons of tree wire and spacer cable to bare designs, no data was found in the literature on the relative tree fault susceptibility for various conductor orientations. As future research work, it would be worthwhile to compile data on tree fault rates for different design configurations (crossarm, armless vertical, and armless candlestick approaches) to see if one is better than the others. It is possible that a widely spaced triangular conductor orientation, such as candlestick, but with much more spacing, could offer improved resistance to certain types of tree faults when using bare conductor. Figure 4-25 compares conventional designs to the proposed widely spaced triangular orientation. Another option is to use covered conductors on just the center phase on three-phase construction, since a phase-to-phase tree contact will normally involve the center phase. That would help with both tree faults and with conductor slapping faults. And by using only one covered phase, the extra cost of covered conductors is reduced, the structure ice and wind loading is reduced, and the probability of a burndown is diminished. The middle phase conductor could also be made larger to reduce the chance of burndowns.

Goodfellow, J., "Understanding the Way Trees Cause Outages," 2000. http://www.eciconsulting.com/images/wnew/trees.pdf.

84

4-48

Construction Issues
Conventional Designs
Horizontal Crossarm Candlestick Armless vertical

Unconventional Wide spaced triangular arrangement using cross arm

Figure 4-25 Example of a Non-Conventional Design That Might be More Resistant to Certain Tree Faults

Another type of tree problem involves a limb or tree that has fallen or blown onto the line in such a manner that it pushes two or more conductors together. This results in a direct contact fault between the wires if they are bare conductors. There is little that can be done to solve this type of fault if a big tree or limb hits a bare conductor line. The weight of a large tree limb and the forces involved are simply too great for the tensions in typical conductors to prevent such contact. For smaller trees and medium size limbs, there is a chance that use of stronger support arrangements, more spacing, and more conductor tension can help reduce some of those faults. Covered conductors or spacer cables could help reduce such scenarios. Faults Due to Wind-Induced Conductor Motion A strong wind alone, without any tree debris, can cause faults by forcing conductors into momentary contact. Aeolian vibrations and galloping effects are types of conductor movement and oscillations that can lead to faults. While these effects may be less critical on distribution systems than on transmission systems, they still play a role in many faults. The best approaches for mitigating these problems are very similar to those discussed earlier for short circuit current conductor slapping faults. That is, make sure the spans are not too long, that the conductor clearances are very large (not just NESC minimums), and that sag is limited. Use of spacers and vibration dampers can help in cases where these conditions cant easily be satisfied. The EPRI 85 Transmission Line Reference Book has a complete chapter on wind-induced motions in
EPRI, Transmission Line Reference Book: 115 - 138 kV Compact Line Design, Second ed, Electric Power Research Institute, Palo Alto, California, 1978.
85

4-49

Construction Issues

conductors. The physic and equations discussed there, while focused on compact transmission line designs, can be applied to distribution scale lines since compact transmission is in some ways not that much different. Mitigating Wind and Tree Faults Overall, there is no easy solution to the problem of trees when it comes to overhead lines. Covered conductors and spacer cable can solve some but not all of the problems related to trees. These technologies can be used in heavily treed areas and provide other benefits besides mitigating tree faults. However the user should be aware that covered conductors, whether they are tree wire or spacer cable, have an increased possibility of burndown and so must be more carefully coordinated with overcurrent protection schemes and properly installed. In addition, spacer cable may have a lower critical flashover voltage (which can increase certain types of lightning faults). Covered conductors and spacer cable still wont stop the heavy tree damage that pulls down the line and pole hardware, but they are good solutions for the smaller limbs and branches.

4-50

5
ANIMAL PROTECTION
Animal faults are a source of significant power quality problems. For many utilities, they are the second or third cause of faults on overhead circuits. Many unknown faults are also likely to be animals. Fortunately, with proper construction practices that include animal protection, utilities can greatly reduce animal faults using relatively simple approaches. EPRI has done a great deal of research into animal-caused faults on distribution lines. The major documents that report this research are:

Mitigation of Animal-Caused Outages for Distribution Lines and Substations, EPRI, Palo Alto, CA: 1999. TE-114915. Distribution Wildlife and Pest Control, EPRI, Palo Alto, CA: 2001. 1001883.

Other useful references on animal protection on distribution circuits include: NRECA86, Frazier 87 88 89 and Bonham , Chow and Taylor , and California Energy Commission .

Animal-Fault Basics
Faults caused by animals are often rank as the number one or number two cause of interruptions for utilities (after trees). An animal that bridges the gap between an energized conductor and ground or another energized phase will create a highly-ionized, low-impedance fault-current path. The fault will cause a voltage sag to nearby customers and an interruption to the portion of the circuit covered by the upstream protective device. An animal can cause a temporary fault or a permanent fault. If the animal remains or charred arc path leaves a conducting path, the circuit will not be able to hold voltage, and the fault will be permanent. If the animal gets blown off or falls away, the circuit can be reenergized (a temporary fault). Animal-caused faults are normally phase to ground. A phase-to-phase flashover path is uncommon but can happen; a three-phase flashover is rare. Squirrels cause the most faults on overhead distribution circuits (see Figure 5-1). Squirrels thrive in suburbs and love trees; utilities have noted increases in squirrel faults following development of wooded areas. Squirrels are creatures of habit and tend to repeatedly follow the same paths.

Animal-Caused Outages, National Rural Electric Cooperative Association, 1996. Frazier, S. D. and Bonham, C., Suggested Practices for Reducing Animal-Caused Outages, IEEE Industry Applications Magazine, vol. 2, no. 4, pp. 2531, August 1996. 88 Chow, M. Y., and Taylor, L. S., Analysis and Prevention of Animal-Caused Faults in Power Distribution Systems, IEEE Transactions on Power Delivery, vol. 10, no. 2, pp. 9951001, April 1995. 89 California Energy Commission, Reducing Wildlife Interactions With Electrical Distribution Facilities, 1999.
87

86

5-1

Animal Protection

Squirrels have a need to gnaw, which can be destructive to utility equipment; they will chew secondary wire coverings and aluminum connectors.

Courtesy: Duke Power

Figure 5-1 A Squirrel Flirting With Danger

Different species of squirrels are different lengths. From longest to shortest, the main species are (lengths are nose to tip-of-tail measurements90):

Fox squirrels: 18 to 27 inches (46 to 69 cm) Eastern and western gray squirrels and tassel-eared squirrels: 16 to 20 inches (41 to 51 cm) Red and Douglass pine squirrels: 10 to 15 inches (25 to 38 cm) Northern flying squirrels: 10 to 12 inches (25 to 30 cm) Southern flying squirrels: 8 to 10 inches (20 to 25 cm)

While we can see that the longest fox squirrels can stretch more than two feet, separations between energized conductors of 18 to 24 inches (46 to 61 cm) are normally enough to prevent most squirrel-caused outages. Spacings less than 10 inches (25 cm) are extremely prone to squirrelsacross an unprotected bushing for example. While squirrels are the most common cause of faults, several other climbing animals can cause faults on overhead circuits, including raccoons, rats, cats, and snakes. While some of these animals are longer than squirrels, the protective measures for squirrels also protects against most faults from other climbing animals.

Jackson, J. J., Tree Squirrels, Cooperative Extension Division, Institute of Agriculture and Natural Resources, University of Nebraska - Lincoln, United States Department of Agriculture, 1994.

90

5-2

Animal Protection

Fox squirrel: dark Tassel-eared squirrel: light

Eastern gray squirrel: dark Western gray squirrel: light

Red squirrel: dark Douglass pine squirrel: light Figure 5-2 Ranges of Squirrel Species

Northern flying squirrel: dark Southern flying squirrel: light

Source: Jackson
90

Common birds (including starlings and blackbirds) rank second behind squirrels as far as the number of animal-caused interruptions caused on overhead distribution circuits 91 92 (EPRI TE-114915 and Frazier and Bonham ). Birds normally cause faults in much the same way as squirrelsby bridging the gap across locations with tight spacings: unprotected bushings or surge arresters, for example. Birds use utility equipment to perch on and for nests (Figure 53). Many of the same protective measures that protect against squirrel-caused faults will also protect against bird-caused faults. Other ways that birds can cause faults include bird-dropping contamination on insulators and woodpecker damage to poles. Large groups of flocking birds (like starlings or crows) can cause a conductor to swing into another conductor when the flock suddenly leaves the conductor. Many large birds, including eagles, hawks, owls, and herons are
Mitigation of Animal-Caused Outages for Distribution Lines and Substations, EPRI, Palo Alto, CA: 1999. TE114915. 92 Frazier, S. D. and Bonham, C., Suggested Practices for Reducing Animal-Caused Outages, IEEE Industry Applications Magazine, vol. 2, no. 4, pp. 2531, August 1996.
91

5-3

Animal Protection

wide enough to span the normal phase-to-phase separations on overhead distribution circuits. Because of this, utilities need to take extra measures to protect these birds from electrocution. As such, it is a matter of protecting endangered animals; in most cases, the power quality impacts of large-raptor contacts are small relative to other animal contacts. Because the power-quality impacts of large-raptor contacts are small, we do not cover that topic in this report.

Courtesy: Duke Power

Figure 5-3 Nest in a Transformer

Most animal-caused faults occur in fair weather. Chow and Taylor93 found that over 85% of animal faults occurred during fair weather on the Duke Power system. They also found that more than half of the animal faults occurred during the morning, and few occurred during the evening or late at night. Squirrels sleep at night and are most active in the morning, when they are looking for food. Some utilities also experience seasonal variations in weather patterns, because animals are less active during the cold of winter. Because animal-caused faults are normally during fair weather, the power-quality perception of these faults is heightened. Utility customers normally expect disturbances during poor weather

Chow, M. Y., and Taylor, L. S., Analysis and Prevention of Animal-Caused Faults in Power Distribution Systems, IEEE Transactions on Power Delivery, vol. 10, no. 2, pp. 9951001, April 1995.

93

5-4

Animal Protection

but not in good weather. When a customer loses a critical process due to a utility interruption or voltage sag, and it is not stormy, the customer is more likely to complain and seek compensation. The types of animals causing faults vary considerably by region, and there is also significant variation within a region. Animal faults also ebb and flow with animal populations. Animal population data can be used as one way to determine if unknown faults are really being caused by certain animals. The patterns of animal-caused faults have been used to classify unknown faults. Chow et. al.94 developed a classification routine to identify animal-caused faults based on the following outage inputs: circuit ID, weather code, time-of-day off, phases affected, and protective device that operated. Animal faults are more likely during fair weather, mornings, only one phase affected, and for a transformer or tap fuse. These same classification strategies can be used to estimate how many of the unknown faults are actually animals. EPRI surveyed utilities on animal faults and animal protective measures . Out of 84 respondents, 77% had some sort of structured program to address animal-caused interruptions. 78% of animalcaused outages were attributed to overhead distribution with the remainder split almost evenly between underground distribution, substation, and transmission. The EPRI survey also points out where most animal-caused outages occurat equipment poles, where phase-to-ground spacings are tight. Figure 5-4 shows that most problem areas are at equipment poles. Transformers are by far the most widely found pole-mounted equipment on a distribution system. Unprotected transformer poles normally have many locations susceptible to animal faults: across the transformer bushing, across a surge arrester, and from a jumper to the transformer tank. Riser pole installations, regulators, reclosers, and capacitor banks all have susceptible flashover paths unless utilities employ protective measures. Poles without equipment are much less susceptible to animal faults (exceptions are poles with grounded guys or ground wires near phase conductors).
95

Chow, M. Y., and Taylor, L. S., A Novel Approach for Distribution Fault Analysis, IEEE Transactions on Power Delivery, vol. 8, no. 4, pp. 18829, October 1993. 95 Mitigation of Animal-Caused Outages for Distribution Lines and Substations, EPRI, Palo Alto, CA: 1999. TE114915.

94

5-5

Animal Protection
Transformers Phasetoneutral contacts Jumper wires contacts Reclosers Surge arresters Cutouts Phasetophase conductor contacts Potheads at riser locations Poletop ground wires Regulators Grounded poletop hardware Insulator contamination Switchgear Breakers Conductor jumping Capacitors Sectionalizers Unknown Structural failure 0 10 20 30 Percentage of utilities listing the item as a topfive outagefrequency problem
96

Data source: EPRI TE-114915

Figure 5-4 Overhead Distribution Points Most Susceptible to Animal-Caused Faults

Animal Guards, Wire Coverings, and Other Protective Equipment

Of the major causes of faultsanimals, trees, lightning, and equipment failuresanimal faults are the most easily prevented. The two main ways to protect equipment against animals (particularly squirrels and birds) are:

Bushing protectors Covered lead wires

Bushing protectors and covered lead wires are inexpensive if installed with equipment (but relatively expensive to retrofit). A varieties of bushing guards are available (see Figure 5-5). These can be used for bushings and surge arresters found on transformers, capacitors, regulators, reclosers, and sectionalizers. Of those locations, transformers are by far the most common. Properly applied bushing guards in conjunction with covered jumper wires can effectively prevent most animal-caused faults. For example, in one Duke Power circuit, animal guard
Mitigation of Animal-Caused Outages for Distribution Lines and Substations, EPRI, Palo Alto, CA: 1999. TE114915.
96

5-6

Animal Protection

application on all distribution transformers on the circuit reduced animal faults from 12 per year to an average of 1.5 per year (Chow and Taylor97). In Lincoln, Nebraska, application of animal guards on all 13,000 transformers reduced the cost of animal-caused faults by 78% (Hamilton et al.98).

Courtesy: Duke Power

Figure 5-5 Various Animal Guards

Animal guards and covered jumper wires prevent most animal flashovers at equipment poles by physically covering energized conductors. See Figure 5-6 for examples of installations. The insulation is not fully rated and provides no degree of safety for line workers (animal guards have no voltage rating, and covered jumper wire is normally 600-V class insulation). The insulation is normally enough to prevent a full flashover across an animal for a momentary contact. For retrofit situations, split-seam insulation covers for jumper wires are available.

Chow, M. Y., and Taylor, L. S., Analysis and Prevention of Animal-Caused Faults in Power Distribution Systems, IEEE Transactions on Power Delivery, vol. 10, no. 2, pp. 9951001, April 1995. 98 Hamilton, J. C., Johnson, R. J., Case, R. M., and Riley, M. W., Assessment of Squirrel-Caused Power Outages, in Annotated Bibliography of Animal Interactions with Utility Structures and Techniques for Problem Mitigation. Vertebrate pest control and management materials., vol. 6, K. A. Fagerstone and R. D. Curnow, Ed. Philadelphia, Pennsylvania, USA: American Society for Testing and Materials, 1989, pp. 3440. ASTM STP 1055.

97

5-7

Animal Protection

Source: EPRI 1001883

99

Figure 5-6 Examples of Installations With Animal Guards and Insulated Jumpers

The EPRI survey found that bushing covers and insulated jumper wires were the most commonly 100 used animal-protective measures employed, as shown in Table 5-1 (EPRI TE-114915 ). Both had an average rating of very good for mitigation success. Bushing covers only rated good for long-term durability. Utilities also use a variety of other measures to combat animals. For this report, we will mainly consider the most common protective measures along with construction practices to reduce animal-caused faults.

Distribution Wildlife and Pest Control, EPRI, Palo Alto, CA: 2001. 1001883. Mitigation of Animal-Caused Outages for Distribution Lines and Substations, EPRI, Palo Alto, CA: 1999. TE114915.
100

99

5-8

Animal Protection

Table 5-1 Surveyed Use of Animal-Mitigating Measures on Overhead Distribution Average Grade for Long-Term Durability
C B B B B B C B B C B B B C

Percentage of Average Grade for Mitigation Utilities Using Success the Method
Insulating Measures Insulated bushing covers Insulated jumper/stinger wires Increased 60 conductor separation for birds-of-prey Wood/fiberglass insulated equipment mounts Extended length deadend insulators to provide additional clearance Insulated bird-of-prey guards Insulating tape Insulated primary wire Heat-shrink insulation material Insulated pole wrapping mastic Silicon insulator coating to reduce bird contamination flashovers Insulating paint Stirrup covers Insulating spray Perching/Nesting Management Triangle perch guards Artificial nesting platforms Artificial elevated bird perches Grounding Solutions Eliminating pole-top ground wires Ground wire molding or insulated ground wires Gapping pole-top ground wires Barriers/Spikes Electrostatic guards Other type climbing guards Metal needle wire guards/spikes Plastic bird spikes Insulated disk barriers Rotating tube barriers for conductors Other Mitigating Measures Fake predators (plastic owl, hawk, etc.) 6 C 17 10 8 8 7 5 B C B C C C 20 12 11 B B C 15 15 11 B B B 64 56 25 20 18 15 14 13 12 10 4 2 2 1 B B B B B B B B B C B B B C

B B B

B C C

B B B C C B

A = excellent; B = very good; C = good; D = poor; F = failed. Data source: EPRI TE-114915101

Mitigation of Animal-Caused Outages for Distribution Lines and Substations, EPRI, Palo Alto, CA: 1999. TE114915.

101

5-9

Animal Protection

Proper installation is important for bushing guards. The guard should be placed securely around the bushing and locked into place according to the manufacturers directions. Height placement is also important (see Figure 5-7). For bushing protectors, have crews leave some room between the bottom of the bushing protector and the tank. Animal guards are not fully rated insulators they can track and flash over, so we do not want the animal guard to fully bypass the insulator.

Source: EPRI 1001883

102

Figure 5-7 Animal Guard Installation

Guard applied too low on the bushing Arrester cap coming loose

Figure 5-8 Examples of Poor Animal Guard Installations

102

Distribution Wildlife and Pest Control, EPRI, Palo Alto, CA: 2001. 1001883.

5-10

Animal Protection

The most common mistake is leaving some energized pathways exposed. Crews should cover every bushing and arrester and use insulated leads on all jumpers. In a survey of 253 poles that should have had animal protection, Pacific Gas & Electric (PG&E) found that 165 poles (65%) 103 were found with incomplete or improperly installed devices (California Energy Commission ). The most common problems were missing bushing covers or missing covers on jumpers; see Figure 5-9.

Cover missing Incomplete jumper installation Cover over too many skirts Cover loose Cover above bushing mount Cover cut Perch guard misplaced Electrostatic guard below first skirt Two covers per bushing 0 10 20 30

Percentage of device problems (N=165)

Data source: California Energy Commission
103

Figure 5-9 Problems Noted at Poles With Missing or Improperly Installed Animal Protective Devices

Animal guards are relatively flimsy compared to most other distribution line hardware. When selecting animal guards, choose models that are large enough to sufficiently cover bushings and arresters. The small protectors that just cover the bushing cap can pop off, and determined animals can wedge between the caps and the bushing. Also, choose animal guards tested to withstand degradation from ultraviolet radiation.
104 In an EPRI survey , utilities reported some animal-mitigating measures caused new outage or maintenance problems. The two reported by more than one utility are:

Bushing covers deteriorating and tracking6 utilities reporting the problem Birds probing for insects in bushing covers, resulting in electrocutions/outages2 utilities reporting the problem

California Energy Commission, Reducing Wildlife Interactions With Electrical Distribution Facilities, 1999. Mitigation of Animal-Caused Outages for Distribution Lines and Substations, EPRI, Palo Alto, CA: 1999. TE114915.
104

103

5-11

Animal Protection

Other single utilities also reported the following problems:

Bushing covers hiding loose connections. Bushing covers hiding lightning damaged insulators. Insects nesting in bushing covers. Heat shrink material reducing the effectiveness of infrared inspections. Equipment coated with rubber paint requires additional time to disassemble. Braided insulated jumper wires breaking in the wind. Improper operation of a spark gap arrester due to incorrectly installed transformer animal guards. Transformer animal guard grounding out a high-side bushing.

Animal guards falling off transformer bushings. Birds nesting on bird guards/spikes. Spikes falling off crossarms. Squirrels nesting under transformer animal guards. Difficulty climbing over mastic pole wraps. Difficulty working around animal spikes. Plastic pole wraps coming loose in the wind. Eagle perches falling off mounting brackets. Raptor perch guards collapsing. Substation snake fences tripping workers. Sonic distress alarms irritating neighbors.

Of all of the problems, the deterioration of bushing covers was the most widely cited in the EPRI survey. Figure 5-10 shows an example of a deteriorated animal guard. Along those lines, four of the utilities in the EPRI survey would like to see products with more resistance to ultraviolet degradation.

5-12

Animal Protection

Source: EPRI TE-114915

105

Figure 5-10 Example of a Deteriorated Animal Guard Compared to a New One

Exposure to ultraviolet radiation causes depolymerization in polymer materials (the polymer bonds break down). The effect of the material depolymerization is a roughening of the material surface, which is often accompanied by cracking and pitting of the surface. The roughened surface may now be more susceptible to pollutant collection and thus more susceptible to dryband arcing and further deterioration. Exposure to ultraviolet radiation is primarily achieved through exposure to sunlight (of which ultraviolet radiation is a component), although ultraviolet radiation is also present in corona discharge, which can create a localized rapid ageing effect. Of the 253 poles examined by Pacific Gas & Electric (PG&E), 80 poles (32%) were found with degraded devices (California Energy Commission106). The type of degradation observed, ranked from least to most, was discoloration (for example, ultra-violet light damage), black traces, tracking and/or erosion, tearing (caused by wear), and deformation. PG&E anticipated that those devices showing discoloration or black traces would have a greater likelihood of performing as they were designed, while those devices showing obvious tracking/erosion, tearing, or deformation would have lost some of their designed functionality. They grouped the results into Classes A and B to represent these less severe (discoloration or black traces) and more severe (tracking/erosion, torn, or deformation) forms of degradation, respectively:

Class ADegradation that is of a lesser degree such as discoloration or black traces that will not likely affect performance.

Mitigation of Animal-Caused Outages for Distribution Lines and Substations, EPRI, Palo Alto, CA: 1999. TE114915. 106 California Energy Commission, Reducing Wildlife Interactions With Electrical Distribution Facilities, 1999.

105

5-13

Animal Protection

Class BDegradation of a greater degree which will likely result in reduced performance such as tears, signs of tracking/erosion, or deformation.

This ranking of degradation severity is based only on what we could observe of the devices condition from the ground. A closer examination could reveal other clues that would indicate greater or lesser degradation. For example, tracking could occur on the inside surface of a device that only appeared discolored on the outside surface. Table 2 summarizes the extent of deterioration on the poles surveyed by PG&E. Of the 80 poles with degraded devices, 43% were exposed to heavy automobile exhaust, and all but one was fully exposed to sunlight (no environmental shielding). Also, most (62%) were in residential areas, and 14% were in agricultural areas.
Table 5-2 Deterioration of Animal-Protective Devices Degradation Class A Class B Number of Poles 69 (27%) 28 (11%) Number of Degraded Devices 91 37

Survey of 253 pole locations Source: California Energy Commission107

PG&E reported that its tests in 1997 found that PVC products perform poorly compared to similar products made from other base materials, such as polypropylene copolymers, ethylene propylene diene methylene (EPDM), and silicone rubber108. Newer products are expected to resist degradation better than older units. The materials are more resistant to degradation from ultraviolet radiation. While there is no standard test requirement for animal guards, some new units are tested according to the ASTM G23 accelerated aging test for 500 or 1000 hours. This section has concentrated on bushing guards and wire coveringsthe most common and effective animal-control technologies. Some additional items that also help include:

Trim treesSquirrels get to utility equipment via trees (pole climbing is less common). If trees are kept away from lines, utility equipment is less attractive. Good outage trackingMany outages are repeated, so a good outage-tracking system can pinpoint hotspots to identify where to target maintenance. Identify animalIf outages are tracked by animal, it is easier to identify proper solutions.

Animal guards and wire coverings help with birds as well as with squirrels and other climbing animals. Additionally, some bird-specific practices include:

107 108

Get rid of nests.

California Energy Commission, Reducing Wildlife Interactions With Electrical Distribution Facilities, 1999. California Energy Commission, Reducing Wildlife Interactions With Electrical Distribution Facilities, 1999.

5-14

Animal Protection

Track as a separate category. Remove nearby roosting areas. For woodpecker damage, Frazier and Bonham109 suggested some home remedies. One option was to leave a woodpecker-damaged pole right next to its new replacement. Because woodpeckers are territorial, a woodpecker will think the area has already been claimed. A similar option is to cut a chunk of wood from a woodpecker hole and attach it to a new pole as a way to mark the territory.

Construction Considerations
Animal faults vary by construction habits within a region. Some common problem areas that can lead to high animal fault rates include:

Unprotected transformer bushingsA very common animal fault is across a transformer bushing. Insulating paints are available for transformers, but it degrades quickly. Bushing guards and insulated lead wires offer the best protection. Unprotected arresters (especially polymer)Another common animal fault is across an arrester, especially a tank-mounted arrester. Polymer-housed arresters have more problems than porcelain-housed arresters because they are much shorter. Use animal guards on arresters, especially on polymer-housed arresters. CutoutsCutouts are sometimes installed such that there is a low clearance between a phase conductor and a grounded object. Grounds and guysUninsulated guy wires and ground wires in close proximity to phase conductors can create locations susceptible to animal faults.

Ground Wires Ground wires near phase conductors can create very short phase-to-ground separations. Figure 5-11 shows an example that illustrates the problem. On this pole, a ground wire runs from the neutral to the top of the pole, leaving less than six inches between the phase and the neutral. This is the type of installation that could have repeated animal faults because of the short phaseto-ground separation. The best solution is to remove the ground wire. It is unnecessary in most cases.

Frazier, S. D. and Bonham, C., Suggested Practices for Reducing Animal-Caused Outages, IEEE Industry Applications Magazine, vol. 2, no. 4, pp. 2531, August 1996.

109

5-15

Animal Protection

Ground wire

Courtesy: Duke Power

Figure 5-11 Electrocuted Squirrel and the Ground Wire That Made the Pole Susceptible to Animal Faults

Utilities have at times bonded wood poles and/or crossarms for two reasons:

Lightning damageLightning can damage wood crossarms. Bonding prevents lightning damage to wood. Pole firesPole fires can start at the interfaces between wood and metal. Bonding shorts out the wood and protects against pole fires.

In most cases, neither of these problems justifies grounding wood poles or crossarms. The incidents of problems from pole fires and lightning damage is less than the problems caused by bonding the wood. Bonding the wood results in more animal faults and more lightning-caused flashovers. If historical records show that wood damage is a problem, bonding the insulators (grounding the base of each insulator) protects the wood, but this shorts out the insulation capability provided by the wood and creates locations where animal contacts can occur. Better solutions are surface 5-16

Animal Protection

electrodes fitted near the insulator pin, including wire-wraps, bands, or other metal extensions attached near the insulator in the likely direction of flashover. This local bonding encourages lightning-caused breakdowns near the surface rather than internally. Another option is to leave gaps in bonding wires; this helps for animal-caused faults but not for lightning-caused faults. Preventative measures for lightning damage to wood also reduce the likelihood of pole-top fires. Leakage current arcs at metal-to-wood interfaces start fires (Darveniza110 and Ross111). Local bonding using wire bands or wraps helps prevent pole damage. Bonding bridges the metal-towood contacts where fires are most likely to start. Local bonding is better than completely bonding the insulators because the insulation level is not compromised, and animal protection can be maintained. Grounded guy wires that attach near primary insulators are also susceptible to animal-caused faults. Support Brackets Grounded support brackets used for cutouts or riser poles can make locations susceptible to animal contacts. Cutouts are difficult to adequately protect with guards and covered jumpers. If the cutout is attached to a grounded base, squirrels or other animals can contact an energized end of the cutout and cause a fault.

110 111

Darveniza, M., Electrical Properties of Wood and Line Design, University of Queensland Press, 1980. Ross, P. M., Burning of Wood Structures by Leakage Currents, AIEE Transactions, vol. 66, pp. 27987, 1947.

5-17

Animal Protection

Source: EPRI 1001883

112

Figure 5-12 Example of a Grounded Bracket

Insulated supports provide better animal protection. A variety of wood or fiberglass options are available. See Figure 5-13 for two examples. The metal ends of these brackets should not be grounded.

112

Distribution Wildlife and Pest Control, EPRI, Palo Alto, CA: 2001. 1001883.

5-18

Animal Protection

Source: EPRI 1001883

113

Figure 5-13 Examples of Insulated Supports

Other metal braces and hardware, including crossarm braces, can contribute to animal-caused faults. Wood crossarm braces provide another level of protection against animal-caused faults. Whether doing design standards or upgrading construction, consider that steel and other conducting supports can lead to animal-caused faults. This also applies to steel poles and crossarmsboth can contribute to structures highly susceptible to animal faults. Fusing Fusing can also change the impact of animal-caused faults for faults across a distribution transformer bushing or an arrester. Having an external fuse improves power quality for several reasons:

Fewer customers interruptedOnly the customers on the transformer have an interruption. Shorter interruption durationBecause a local fuse isolates the problem to a small area, crews do not have to patrol circuits to look for problems. The crew can drive right to the problem location. It also helps crews more accurately identify the source of the problem as an animal-caused fault.
Distribution Wildlife and Pest Control, EPRI, Palo Alto, CA: 2001. 1001883.

113

5-19

Animal Protection

Fewer momentary interruptionsA local fuse may reduce momentary interruptions on circuits with fuse saving because a small transformer fuse may clear before the substation circuit breaker trips. Less-severe voltage sagsA transformer fuse will clear faults faster than the next upstream device (a lateral fuse, a recloser, or a circuit breaker), reducing the duration of voltage sags to customers on the system. If the location is near the substation (high fault currents), and the location has a current-limiting fuse, then the current-limiting fuse will reduce the magnitude of the voltage sag.

If the transformer is a completely self-protected transformer (CSP, see Figure 5-14) with an internal fuse, then the tap fuse or upstream circuit breaker or recloser operates, leaving many more customers interrupted, with much more area for crews to patrol. CSPs on the mainline are especially problematic for reliability indices.

Figure 5-14 CSP Transformer Exposing an Entire Circuit Tap to Animal Faults

Line surge arresters are another common overhead-line application that normally does not have a local fuse (see Figure 5-15). Line arresters are arresters applied not to protect any particular piece of equipment but to improve the performance of the line itself. On such installations, it is important to provide adequate animal protections. Another possibility is to consider fusing the arresters for additional protection. 5-20

Animal Protection

Figure 5-15 Line Arresters That Leave the Mainline Exposed to Animal Faults Across the Arresters

Summary of Animal-Protection Guidelines

Most animal-caused faults can be prevented with proper application of animal guards and covered jumpers along with good construction practices that minimize the possibility of animal contacts. Following these practices should eliminate 80 to 90% of animal-caused faults on overhead distribution circuits, but not all can be eliminated. Figure 5-16 shows an example of damage from a nest fire, despite the application of bushing guards, wooden crossarm braces, and relatively wide conductor spacings (although the phase-to-phase spacing between the cutouts on the crossarm is less than it should be).

5-21

Animal Protection

Source: EPRI 1001883

114

Figure 5-16 Pole Charred After a Nest Fire

Consider the following recommendations to reduce animal-caused faults and improve power quality:

Animal guards and protected jumpersUse animal guards and covered jumper wires on new installations. For problem areas, retrofit installations as needed. Animal guard applicationsTrain crews to properly install animal guards. Ensure that guards fit tightly but do not cover too much of the bushing. Animal guard specificationsChoose and specify animal guards appropriately. Use an animal guard that is large enough to adequately cover the top of the bushing. Do not just specify and purchase equipment (transformers, arresters, and so on) with animal protection. A specification of animal protection can lead to manufacturers applying the cheapest animal guards, which may be flimsy or undersized. Ground wiresEliminate unneeded ground wires on poles. Most problematic are ground wires run up near phase conductors. In most cases, bonding of wood members is not needed. Where ground wires are needed near phase conductors, use insulated ground wires. Guy wiresInsulate guy wires. Grounded guy wires near phase insulators create an opportune animal flashover location.
Distribution Wildlife and Pest Control, EPRI, Palo Alto, CA: 2001. 1001883.

114

5-22

Animal Protection

Structure supportsWhere possible, use insulating structure supports rather than grounded structure supports on riser poles and other locations where a grounded support provides a convenient platform for animal contact. Energized phase spacingsWhere possible, separate energized conductors from grounded conductors and other phases. Separations of at least 18 to 24 inches (46 to 61 cm) are normally sufficient; spacings less than 10 inches (25 cm) are very prone to animal faults. Where sufficient air clearances are impossible, use insulated phase and/or ground conductors. Equipment fusingAt locations susceptible to animal faults (mainly equipment poles), consider an external fuse on structures that are upstream of likely animal contacts. Two applications that utilities often leave exposed without a local fuse are line arresters and completely self-protected transformers (CSPs). TargetingTrack animal-caused faults and target appropriately.

For the future, the biggest need for the industry is addressing the deterioration of animal guards. This could include investigation of the aging mechanisms in the polymer materials and development of testing protocols and possibly standardization.

5-23

6
CONSTRUCTION AND MAINTENANCE PROGRAMS TO REDUCE TREE FAULTS
Trees are the source of many power quality issues on distribution and transmission circuits. For many utilities, trees are the number one or number two cause of interruptions. When trees contact utility equipment, damage is often extensive, and repair is expensive and timeconsuming. Most of these vegetation faults are on distribution circuits, but transmission circuits are not immune as highlighted by the famous tree-caused fault on August 14, 2003 on one of FirstEnergys 345-kV lines in Ohio. In addition to long-duration interruptions, the faults from trees cause voltage sags and can cause momentary interruptions. For most utilities, vegetation management is by far the largest maintenance item in the budget. So, in addition to improving power quality, more efficient tree maintenance and more treeresistant designs can also reduce maintenance.

Tree Faults
Contact between trees and electric distribution circuits can create significant power quality problems resulting in momentary interruptions, voltage sags, and flicker. Momentary Interruptions from Trees A question that is difficult to answer is: how often do trees cause momentary interruptions? There is a common perception that trees cause momentary interruptions, especially during storms. Momentaries could occur in the following ways:

Temporary faultsIf trees cause a temporary fault (no permanent damage) from phase to phase or from phase to ground, a circuit breaker or recloser can open and then reclose successfully. Permanent faultsPermanent faults from trees on fused taps will cause momentary interruptions to the circuit if the utility uses fuse saving; the upstream circuit breaker or recloser will try to open before the fuse operates to reduce unnecessary fuse operations.

The second case is predictable. The first case is less predictable, and we are unaware of any tests or monitoring to quantify the number of temporary faults that trees might cause. Two possible scenarios for temporary faults are:

6-1

Construction and Maintenance Programs to Reduce Tree Faults

1. WindWind could push trees into conductors and cause them to slap together. Once the conductors come apart, the insulation is restored, and the breaker or recloser can reclose successfully. 2. RainRain could weigh down tree branches, causing them to sag into a circuit. The wet branches making phase-to-phase contact cause a fault. Then the heat and explosive blast from the fault arc will evaporate and shake water off of the tree branches causing them to rise away from the conductors and restore insulation. As more rain accumulates, the branches can again sag into the conductors. One thing that cannot cause a temporary fault is a tree touching just one phase conductor. Several utilities and researchers have tested a distribution voltage phase in contact with a tree (Germany115, Baltimore Gas & Electric116 , Florida Power Corp117, Texas A&M118, Niagara Mohawk Power Corporation119). In all cases, there may be burning near the contact point, but the contact draws little current, and it does not fault. So if temporary faults occur regularly from trees, they must be across wire-to-wire contacts in close proximity, and not just wire to tree. To explore this further, outage data from a southeastern utility was used to estimate the relationship between feeder lockouts and momentary interruptions. This utility tracks momentary interruptions. Figure 5-36 plots the number of feeder lockouts in a day against the number of feeder momentary events per day on the utilitys system. Were using feeder lockouts as a proxy for tree-caused faults (which were not available in this dataset) under the assumption that days with high numbers of feeder lockouts are during stormy weather, and most lockouts are due to tree faults. Figure 5-36 shows that momentary interruptions generally increase with increasing number of feeder lockouts. But there were days with 5 to 15 lockouts where there were not a lot of momentaries.

Hoffmann, E., Rheinbaben, H. v., and Stosser, B., "Trees in Electrical Contact with High Voltage Lines," CIGRE Session, 22-03, Sept. 1984. 116 Rees, W. T., Birx, T. C., Neal, D. L., Summerson, C. J., Tiburzi, F. L., and Thurber, J. A., "Priority Trimming to Improve Reliability," ISA conference, Halifax, NS, 1994. 117 Williams, C., "Tree Fault Testing on a 12 kV Distribution System," IEEE/PES Winter Power Meeting presentation to the Distribution Subcommittee, 1999. 118 Butler, K. L., Russell, B. D., Benner, C., and Andoh, K., "Characterization of Electrical Incipient Fault Signature Resulting from Tree Contact with Electric Distribution Feeders," IEEE Power Engineering Society Summer Meeting, 1999. 119 Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001.

115

6-2

Construction and Maintenance Programs to Reduce Tree Faults

150
Feeder momentary events

100

50

0 0 5 10 Feeder lockouts 15 20

Figure 6-1 Feeder Lockouts Versus Momentary Events

Lightning and animals are thought to cause most temporary faults. During storms, lightning causes many faults, but animals cause few. To analyze this data in more detail, the data in Figure 5-36 was broken down by month in Figure 5-37. Note that the relationship between momentaries and feeder lockouts greatly increases during the summer months during the peak lightning season for this utility. During September through April when there is little lightning, the relationship is weaker (but its still there). That suggests that if trees regularly cause momentary interruptions, they do so at a rate thats much smaller than lightning-caused faults. This is a rather shaky analysis, so more data or experimental data is recommended before drawing too many conclusions from this analysis.

6-3

Construction and Maintenance Programs to Reduce Tree Faults

0 100 80 60 40 20 2 4 6 8 10 0 2 4 6 8 10

Jan

Feb

Mar

Apr

Feeder momentary events

May

Jun

Jul

Aug

100 80 60 40 20

100 80 60 40 20 0 0 2

Sep

Oct

Nov

Dec

8 10

8 10

Feeder lockouts
Figure 6-2 Feeder Lockouts Versus Momentary Events by Month

Voltage Sags From Trees Tree faults cause voltage sags, as do other types of faults. Tree-caused faults are more likely to cause more severe voltage sags than other faults for two reasons:

Multiple-phase faultsTrees often cause multiple-phase faults, and multiple-phase faults cause more severe voltage sags than single-phase faults. Permanent interruptionsTrees normally cause permanent interruptions. For faults on the mainline, this will cause the circuit breaker or recloser to cycle through its reclose sequence and lock out. The full reclosing sequence causes multiple sag events, and often long-duration sag events (because of the time-delay element of the circuit breaker) to customers on adjacent circuits.

6-4

Construction and Maintenance Programs to Reduce Tree Faults

Voltage Flicker From Trees Can trees cause voltage flicker? Tests by BGE and others have shown that once a limb carbonizes fully, the arc is close to a solid short circuit with no sputtering or significant arc impedance. But before a tree limb fully carbonizes, can limb and branch burning and arcing draw enough current to cause flicker on a circuit? To try and answer this question, we need to examine how much primary current is necessary to cause flicker and what happens to the resistance of tree branches. Small limbs can arc and burn off before the phase-to-phase or phase-to-ground path is fully bridged. This can happen repeatedly. The current drawn is mainly a function of the resistance of the tree branch bridging the gap. Whether this causes flicker or not depends on the stiffness of the utility system at this point. The widely cited GE flicker curve shown in Figure 5-38 can serve as a baseline for judging how severe voltage deviations need to be before utility customers notice light flicker. Because of the arcing nature of the progressing breakdown across a tree limb, we should assume the worst-case frequency on the flicker curve where voltage changes of 0.3% can be troublesome.
6 5 4 3 2 1 0 1 2 5 10 20 30 1 2 5 10 20 30 1 2 5 10 20 DIPS PER HOUR DIPS PER MINUTE DIPS PER SECOND FREQUENCY OF DIPS 30 12 6 3 2 1 30 12 6 3 2 1 . 3 . 2 .05 MINUTES SECONDS TIME BETWEEN DIPS

BORDERLINE OF IRRITATION BORDERLINE OF VISIBILITY OF FLICKER

Figure 6-3 GE Voltage Flicker Curve

The voltage drop from current depends on the system impedance and the current as:
Vdrop R I R + X I X
Eq. 6-1

where R and X are the source resistance and reactance, and IR is the resistive component of the load, and IX is the reactive component of the load. The impedance of an arc is primarily resistive, but it can be highly nonlinear. The arc nonlinearity will create higher frequencies in the current drawn by the arc. The higher frequency current will raise the resistance of the system impedance (wires and cables have higher resistance to higher frequencies). As a first approximation, let us assume that the effective resistance is 6-5

Construction and Maintenance Programs to Reduce Tree Faults

three-times the fundamental-frequency resistance. We also assume that the current is all resistive. Based on these assumptions, Table 5-7 shows some ranges of current necessary to cause voltage flicker that could be observable.
Table 6-1 Current Draw Necessary to Cause Observable Light Flicker Current necessary to cause a 0.3% voltage deviation [amperes] 6.8 3.4 1.7 Tree resistance to draw that current [kiloohms] 1.1 2.1 4.2

Distance of the tree from the substation [miles] 5 10 20 Assuming 7.2 kV line to ground

Effective system resistance [ohms] 3.2 6.4 12.7

A tree in contact with just one phase conductor is unlikely to draw enough current to cause flicker. Several groups have tested or measured such contacts:

Germany Hoffmann et al.120 tested a tree in contact to a 20-kV line, which drew 80 mA initially then rose to 600 mA. They also measured a poplar tree and found the distribution of impedances shown in Figure 5-39 with impedances of 1.8 k/foot (6 k/m) near the base of the tree rising to 25 k/foot (80 k/m) near the top of the tree. The total impedance of this tree is 320 k, which would draw from 0.2 to 0.7 A from a 12.47-kV line. Hoffman also found resistances from the base of the tree to ground of about 300 for soil resistivities near 600 -m. Baltimore Gas and ElectricRees et al.121reported on a sequence of tests on an abandoned distribution tap. Seven tests were performed with different tree species held in contact with a 7.6-kVL-G circuit. Although measurements were difficult because of low readings, BGE calculated amperages from 6 to 41 mA. Two of the tests were done under a constant spray of water that thoroughly saturated the trees and overhead equipment. There was no significant difference when the same species of tree was tested with and without the water spray. Niagara MohawkTested several trees at 7.62 kV from line to groundthe highest was just under 0.5 A (see Table 5-8). Texas A&MThese were the only tests to produce enough current to possibly cause voltage flicker. Butler et al.122 tested two different sized branches in contact with a 7.2-kVL-G distribution circuit. In a test of two trees in contact with branches from pencil thickness to about two inches in diameter, burning and arcing was noted during the tests. The current

Hoffmann, E., Rheinbaben, H. v., and Stosser, B., "Trees in Electrical Contact with High Voltage Lines," CIGRE Session, 22-03, Sept. 1984. 121 Rees, W. T., Birx, T. C., Neal, D. L., Summerson, C. J., Tiburzi, F. L., and Thurber, J. A., "Priority Trimming to Improve Reliability," ISA conference, Halifax, NS, 1994. 122 Butler, K. L., Russell, B. D., Benner, C., and Andoh, K., "Characterization of Electrical Incipient Fault Signature Resulting from Tree Contact with Electric Distribution Feeders," IEEE Power Engineering Society Summer Meeting, 1999.

120

6-6

Construction and Maintenance Programs to Reduce Tree Faults

started and 1.2 A and then climbed to 2.2 A, but it did not fluctuate significantly. In another case with contact to a five-inch diameter branch, the current reached almost 2.5 A but again did not fluctuate. Larger tree limbs have lower resistances, so the worst scenario is a primary conductor solidly contacting a large branch or the main trunk with a low soil resistivity. Even with that, drawing currents high enough to cause flicker is unlikely. The only test that showed current approaching that needed to cause flicker did not fluctuate significantly.
0.18 m 1.8 m 2m 0.3 cm 2.5 cm 4 cm 50 k 150 k 60 k

10.8 m 5m 48 k

20 cm 2m 25 cm 12 k 1 k

Source: Hoffmann, E., Rheinbaben, H. v., and Stosser, B., "Trees in Electrical Contact with High Voltage Lines," CIGRE Session, 22-03, Sept. 1984. Figure 6-4 Resistance Measurements on a Live Poplar Tree Table 6-2 Current Drawn by Several Trees in Contact With 7.62 kV From Line to Ground Location Test 1 Test 2 Test 3 Test 4 Test 5 Test 6 Test 7 Specimen Black Gum Black Gum Black Cherry Black Cherry White Ash Aspen Red Maple Maximum current, mA 92 69 110 100 37 484 125.5

Source: Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001.

6-7

Construction and Maintenance Programs to Reduce Tree Faults

If tree contacts from phase to ground cannot cause flicker, what about contacts from phase to phase or phase to neutral? The voltage gradient is much higher, so such contacts should pull more current. The current is a function of the tree branch resistances. A number of groups have measured tree resistivities:

Defandorf123 measured tree resistivities of a green tulip tree and found resistivities from 50 to 120 -m. He reported that the effective resistivity increased for larger diameter branches and trunks. Hoffmann et al.124reported resistivities of live poplar as having resistivity of 40 to 100 -m. Daily125 measured resistivities of American elms and found values from 37 to 55 -m.

Resistivities can vary significantly by species as shown in Figure 5-40. A one-inch diameter branch with a 50 -m resistivity has a resistance of 30 k/foot (see Table 5-9). To cause flicker, the impedances need to be less than 5000 ohms (per Table 5-7). It appears that branches less than one inch in diameter cannot cause flicker. To get to 5000 ohms across a three-foot gap with oneinch diameter branches, it would take 20 branches in parallel. Although 20 branches in parallel is easily possible on a heavily treed circuit, it is the sputtering of an individual branch that would lead to flicker. If each of 20 branches fluctuated in impedance, the randomness would cancel the flickering effect: all 20 paths would have to fluctuate together.
Ponderosa Pine Paper Birch Green Ash Sycamore Big Leaf Maple Douglas Fir Sugar Maple Eastern White Pine Quaking Aspen Pacific Black Cottonwood Red Alder Red Maple Silver Maple Siberian Elm Weeping Willow 0 500 1000 Tree resistivity, ohmm 1500 2000

Source: Appelt, P. J. and Goodfellow, J. W., "Research on how trees cause interruptions - applications to vegetation management," Rural Electric Power Conference, 2004. Figure 6-5 Tree Resistivities

Defandorf, F. M., "Electrical Resistance to the Earth of a Live Tree," AIEE Transactions, vol. 75, pp. 936-41, Oct. 1956. 124 Hoffmann, E., Rheinbaben, H. v., and Stosser, B., "Trees in Electrical Contact with High Voltage Lines," CIGRE Session, 22-03, Sept. 1984. 125 Daily, W. K., "Engineering Justification for Tree Trimming [Power System Maintenance]," IEEE Transactions on Power Delivery, vol. 14, no. 4, pp. 1511-8, October 1999.

123

6-8

Construction and Maintenance Programs to Reduce Tree Faults Table 6-3 Tree Branch Resistance Diameter [inches] 0.5 1.0 5.0 Resistance [ohms/foot] 120,000 30,000 1,200

On higher voltage distribution circuits and on subtransmission circuits, the voltage gradient is higher across tree limbs. This will pull more current through tree limbs, and they will progress to a full-fledged fault faster. Limbs are less likely to burn clear before they fully fault. Also, because the source impedances are stiffer on higher-voltage circuits, it takes more current to cause flicker. Based on this analysis, voltage flicker from trees is unlikely. During a severe storm, customers may interpret the voltage sags from tree-caused faults as flicker, but flicker from trees contacting phase conductors or from branch burning is unlikely. Previous Work A great deal of work on tree faults has been done since 1990 that should help utilities design more tree-resistant structures and optimize vegetation management. Rees et al.126 reported on pioneering investigations by Baltimore Gas and Electric (BGE) that focused on a seven-year outage-review study and staged tree-fault testing:

Outage reviewBased on a review of over 3000 tree-related outages, BGE concluded that 98% of tree-caused outages were from trees or tree parts falling on lines. Field testsIn tests of tree limbs across line-to-ground voltage (7.6 kV), BGE found that the tree limb did not immediately fault. Burning and arcing starts at the ends and moves inward. The burning carbonizes a highly conductive path. Prior to full development of the carbon path, the tree limb remains a high impedance. Once the carbon path fully develops, the impedance drops to near zero and lead to a fully bolted fault. BGE also staged tests with a live 7.6-kV (line to ground) circuit pulled into contact with a tree from line to the ground. The circuit did not fault and drew currents of less than one ampere in all tests.

Based on these findings, BGE switched to a more prioritized vegetation management program on its distribution system (13.2 kV mainly). They focused less on trimming for natural growth and attempted to remove overhangs where possible. They implemented a three-year maintenance cycle on the three-phase system, and delegated the one and two-phase system to trim only as necessary. BGE crews also removed hazard trees. On the 34.5-kV subtransmission system, they moved to a biannual inspection program with the goal of achieving reliability approaching that of their transmission lines.
Rees, W. T., Birx, T. C., Neal, D. L., Summerson, C. J., Tiburzi, F. L., and Thurber, J. A., "Priority Trimming to Improve Reliability," ISA conference, Halifax, NS, 1994.
126

6-9

Construction and Maintenance Programs to Reduce Tree Faults

Duke Power has done considerable work on using their outage database as a resource to learn 127 about tree-caused faults to help guide vegetation management. Chow and Taylor developed a strategy to analyze Duke Powers outage database to learn more about specific causes of faults. They found the following trends:

WeatherWhen looking at the likelihood of tree-caused faults, weather strongly affects tree faults, especially wind but also rain, snow, and ice. Season and time of dayThe most tree faults occurred during summer and the least occurred during the winter. More tree faults occurred during the afternoon and evening. Tree faults were not greatly influenced by the day of the week. Phasing and protective deviceMultiple-phase faults are more likely to be caused by trees. Along these same lines, lockouts of a substation circuit breaker or a line recloser were more likely to be caused by trees than were operations of other protective devices.

More recent work reported by Xu et al.128 found many of the same trends and extended the concept to include a statistical regression analysis to identify the variables that most influence the likelihood of a tree-caused interruption. They found that weather, time of day, and protective device were most statistically significant indicators of the likelihood of a tree-caused interruption. Niagara Mohawk (a National Grid company) used several investigations to restructure its vegetation management programs for distribution systems129,130. Niagara Mohawk used its outage cause codes to categorize the source of tree-caused interruptions, and also used sample studies to provide more in-depth details on tree-caused interruptions. Niagara Mohawk also staged live tests with trees in contact with distribution lines to learn more about momentary interruptions. They also reviewed a sample of tree-caused outages for tree defects. Based on these results, Niagara Mohawk modified its program to target the worst circuits, the 13.2-kV system, the circuit backbone, and danger-tree removal. With Allegany Power, Niagara Mohawk, and Portland General sponsoring the work, Environmental Consultants Inc. (ECI) performed a large number of tests on various species and 131,132,133 . ECI generally found similar results sizes of tree branches to extend the BGE testing work to the BGE tests: a phase-to-phase or phase-to-neutral contact was required to cause a fault. They also found that the probability of a fault and the time to fault was a function of the voltage

Chow, M. Y. and Taylor, L. S., "A Novel Approach for Distribution Fault Analysis," IEEE Transactions on Power Delivery, vol. 8, no. 4, pp. 1882-9, October 1993. 128 Xu, L., Chow, M.-Y., and Taylor, L. S., "Analysis of Tree-Caused Faults in Power Distribution Systems," 35th North American Power Symposium, University of Missouri-Rolla in Rolla, Missouri, October 20-21, 2003. 129 Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001. 130 Finch, K., "Understanding Line Clearance and Tree Caused Outages," EEI Natural Resources Workshop, April 1, 2003a. 131 Appelt, P. J. and Goodfellow, J. W., "Research on how trees cause interruptions - applications to vegetation management," Rural Electric Power Conference, 2004. 132 Finch, K., "Understanding Line Clearance and Tree Caused Outages," EEI Natural Resources Workshop, April 1, 2003a. 133 Goodfellow, J., "Understanding the Way Trees Cause Outages," 2000. http://www.eciconsulting.com/images/wnew/trees.pdf.

127

6-10

Construction and Maintenance Programs to Reduce Tree Faults

gradient and to some degree a function of the species and moisture content. In addition, ECI surveyed utilities on their tree-caused outages to identify the sources of these outages. Tree Fault Case Studies Faults caused by trees generally occur from a handful of conditions: 1. Falling trees or major limbs knock down poles or break pole hardware 2. Tree branches blown by the wind push conductors together 3. A branch falls across the wires and forms a bridge from conductor to conductor 4. Natural tree growth causes a bridge across conductors The results from several utilities outlined below show that broken tree branches or falling trees account for the majority of interruptions. Growth only accounts for a small percentage of interruptions. Baltimore Gas and Electric The survey of Rees et al.134 of over 3000 tree-related outages over seven years found the following breakdown of tree-caused outages:

75% were caused by dead shorts across a limited distance. 23% were from mechanical damage to utility equipment. Less than 2% were due to natural growth or burning branch tips beneath the lines.

BGE concluded that 98% of tree-caused outages were from trees or tree parts falling on lines. Eastern Utilities Associates
135 Simpson reported on a survey of tree-caused faults at Eastern Utilities Associates (a small utility in Massachusetts that is now a part of National Grid). The main results were that treecaused outages broke down as follows:

63% from broken branches 11% from falling trees 2% from tree growth

Rees, W. T., Birx, T. C., Neal, D. L., Summerson, C. J., Tiburzi, F. L., and Thurber, J. A., "Priority Trimming to Improve Reliability," ISA conference, Halifax, NS, 1994. 135 Simpson, P., "EUA's Dual Approach Reduces Tree-Caused Outages," Transmission & Distribution World, pp. 228, August 1997.

134

6-11

Construction and Maintenance Programs to Reduce Tree Faults

BC Hydro EPRI 1008480136 documents surveys by BC Hydro of their tree-related interruptions that were as follows:

70% from tree failure 18% from branch failure 12% from growth

Niagara Mohawk Finch137 reported on a survey that Niagara Mohawk Power Corporation (now a part of National Grid) performed in 2000. Significant results were:

86% of permanent tree-faults were from outside of the trim zone (+/- 10 feet). Growth only accounted for 14% of outages (Figure 5-17), and Finch also reported that most of these were outages on services. In a sample of 250 tree-caused outages from 1995, 36% were from dead trees, and 64% were from live trees. In this sample, 75% came from outside the trim zone, and 62% were caused by a broken trunk or branch.

Tree fell Small limb Large limb Growth

10

20

30

40

Percent of permanent tree faults by cause

Data source: Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001.

Figure 6-6 Niagara Mohawk Survey of Tree Outage Causes

Duke Power Taylor138 reported on a 1995 sample survey of tree-caused outages with the following results:

EPRI 1008480, Electric Distribution Hazard Tree Risk Reduction Strategies, Electric Power Research Institute, Palo Alto, CA, 2004. 137 Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001.

136

6-12

Construction and Maintenance Programs to Reduce Tree Faults

73% of tree outages occurred when an entire tree fell on the line. 86% of these were from outside of a 30 foot ROW. Dead limbs or trees caused 45% of tree outages.

In addition, Duke Powers investigations of the sample outage set found that 25% of outages reported as tree-caused were not caused by trees. This highlights the importance of good outage code recording when investigating fault causes. Environmental Consultants Inc. Finch reported on results of a 1995 Environmental Consultants Inc. (ECI) survey of over 20 utilities and a total of 2328 tree outages. The ECI survey found that tree failures and limbs caused the most tree outages (see Figure 5-18).
Limb Growth Trunk Uproot Cut Other 0 5 10 15 20 25 30
139

Percent of treecaused faults

Data source: Finch139 Figure 6-7 ECI Survey of Tree Outage Causes

Southeastern US Utility One southeastern US utility uses detailed tree outage codes, allowing them to target causes more precisely. See Table 5-3 for their breakdown of tree-faults and their impact on outages. Note that trees falling (whether from inside or outside of the right of way) cause a much larger impact on the customer minutes of interruption relative to the actual number of outages. Likewise, vines and tree growth have relatively less impact on outage duration.

Taylor, L., "The Illusion of Knowledge," Southeastern Electric Exchange Power Quality and Reliability Committee, Dallas, TX, 2003. 139 Finch, K., "Understanding Line Clearance and Tree Caused Outages," EEI Natural Resources Workshop, April 1, 2003a.

138

6-13

Construction and Maintenance Programs to Reduce Tree Faults Table 6-4 Percentage of Tree Faults in Each Category Outages Tree Outside Right of Way (Fall/Lean On Primary) Tree/Limb Growth Limb Fell from Outside Right of Way Tree Inside Right of Way (Fall/Lean On Primary) Vines Limb Fell from Inside Right of Way Tree on Multiplex Cable or Open Wire Secondary Source: Southeastern US utility, 2003 2004 CI = Customer interruptions CMI = Customer minutes of interruptions 26.0 21.1 18.0 12.6 10.0 8.7 3.6 CI 37.2 14.4 20.1 14.8 3.6 9.8 0.2 CMI 42.5 13.3 18.1 15.2 3.1 7.5 0.2

Characteristics of Tree Faults and Interruptions The outage database is a prime source of information for a utility in targeting tree-caused faults. Tree-caused faults are the most significant contributor to reliability indices. Figure 5-19 shows a breakdown by outage cause code for a utility in the northeast US over a tenyear period (with major storms excluded).
Tree Contact Lightning Cable Fault Animal Contact Other Power Supply 0 5 10 15 20 25 30

Percent of SAIDI Tree Contact Lightning Cable Fault Animal Contact Other Power Supply 0 5 10 15 20 25 30

Percent of SAIFI

Figure 6-8 Breakdown of Reliability Indices for a Northeastern Utility

Analyzing cause codes along with weather codes can reveal significant trends for tree faults. For one northeastern US utility, Figure 5-20 shows the breakdown of outages by weather 6-14

Construction and Maintenance Programs to Reduce Tree Faults

classification. For this utility, Tree Contact leads all other outage categories during stormy weather; and it especially stands out during wind and ice/snow. This highlights the fact that treecaused faults increase rapidly during storms. Figure 5-21 shows the same data but rearranged to focus on comparing tree faults to other common outages caused by weather.
Ice/Snow
Tree Contact Lightning Animal Contact Conductor Off Pin Cutout Fuse Vehicle Accident Cable Fault Wire Shorted Other Insulator Transformer Lightning Arrestor Conversion Hot Line Clamp Clearance

Major Storm

Wind

Rain
Tree Contact Lightning Animal Contact Conductor Off Pin Cutout Fuse Vehicle Accident Cable Fault Wire Shorted Other Insulator Transformer Lightning Arrestor Conversion Hot Line Clamp Clearance

Other

Rain/Wind

Cold
Tree Contact Lightning Animal Contact Conductor Off Pin Cutout Fuse Vehicle Accident Cable Fault Wire Shorted Other Insulator Transformer Lightning Arrestor Conversion Hot Line Clamp Clearance 0 10 20 30 40 50 0

Hot

Fair

10 20 30 40 50

10 20 30 40 50

Percent of outages during the given weather condition

Figure 6-9 Percentage of Interruptions by Cause for the Given Weather for a Northeastern Utility

6-15

Construction and Maintenance Programs to Reduce Tree Faults

Tree Contact
Wind Major Storm Ice/Snow Rain/Wind Other Rain Fair Hot Cold

Lightning

Animal Contact

Cable Fault

0 10 20 30 40 50 60 0 10 20 30 40 50 60 0 10 20 30 40 50 60 0 10 20 30 40 50 60

Percent likelihood of the cause given the specified weather

Figure 6-10 Likelihood of the Given Cause for the Specified Weather for a Northeastern Utility

For many utilities, tree-caused interruptions are the main cause of interruptions during major storms or major events (the IEEE terminology). Duke Power found that trees account for 81% of their outages during major event days140. The data for a northeastern US utility in Figure 5-22 also shows tree-contact as the main cause during major events (with lightning a close second). The effect of trees during major events depends on the dominant storm patterns and the utilitys definition of a major event.

140

Taylor, L., 2004a. Email to the IEEE Task Force on Outage Reporting.

6-16

Construction and Maintenance Programs to Reduce Tree Faults

During major storms SAIDI

Tree Contact Lightning Power Supply Cable Fault Other Wire Shorted Line Fuse Overload Insulator

During major storms SAIFI

Major storms excluded SAIDI

Tree Contact Lightning Power Supply Cable Fault Other Wire Shorted Line Fuse Overload Insulator

Major storms excluded SAIFI

Major storms included SAIDI

Tree Contact Lightning Power Supply Cable Fault Other Wire Shorted Line Fuse Overload Insulator 0 10 20 30 40 0

Major storms included SAIFI

10

20

30

40

Percent
Figure 6-11 Interruption Indices by Cause During Major Storms and With and Without Major Storms

6-17

Construction and Maintenance Programs to Reduce Tree Faults

Being highly correlated with weather, season has a large impact on tree-fault rates. Duke Power has the most tree-caused faults during the summer and the least during fall141,142. Another southeastern US utility also has the most outages during the summer as shown in Figure 5-23 as does a northeastern US utility as shown in Figure 5-24.
Outages
20

CI

CMI

Percent in the given month

15

10

Feb Apr Jun Aug Oct Dec

Feb Apr Jun Aug Oct Dec

Feb Apr Jun Aug Oct Dec

Source: Southeastern US utility, 2003 2004 Major Events Included CI = Customer interruptions CMI = Customer minutes of interruptions Figure 6-12 Tree-Caused Outage Impacts by Month

Chow, M. Y. and Taylor, L. S., "A Novel Approach for Distribution Fault Analysis," IEEE Transactions on Power Delivery, vol. 8, no. 4, pp. 1882-9, October 1993. 142 Xu, L., Chow, M.-Y., and Taylor, L. S., "Analysis of Tree-Caused Faults in Power Distribution Systems," 35th North American Power Symposium, University of Missouri-Rolla in Rolla, Missouri, October 20-21, 2003.

141

6-18

Construction and Maintenance Programs to Reduce Tree Faults

25 20 15 10 5 0 25 20 15 10 5 0

SAIFI

SAIDI

Outages

Percent of yearly total

Treecaused SAIFI

Treecaused SAIDI

Treecaused Outages

Feb Apr Jun Aug Oct Dec

Feb Apr Jun Aug Oct Dec

Feb Apr Jun Aug Oct Dec

Figure 6-13 Interruptions by Month for One Northeastern Utility (Major Events Included)

Interruptions on the mainline affect the most customers. Figure 5-25 shows the breakdown of tree-caused SAIFI for one northeastern US utility. This utility had almost 60% of tree customer 143 interruptions from breaker lockouts. Duke Power also found that of circuit breaker and recloser lockouts, trees caused 35 to 50% of the lockouts, which is over twice the rate of all tree outage events (trees cause 15 to 20% of all of Dukes outages).

Chow, M. Y. and Taylor, L. S., "A Novel Approach for Distribution Fault Analysis," IEEE Transactions on Power Delivery, vol. 8, no. 4, pp. 1882-9, October 1993. 143 Xu, L., Chow, M.-Y., and Taylor, L. S., "Analysis of Tree-Caused Faults in Power Distribution Systems," 35th North American Power Symposium, University of Missouri-Rolla in Rolla, Missouri, October 20-21, 2003.

143

6-19

Construction and Maintenance Programs to Reduce Tree Faults

Circuit Breaker Recloser Line Fuse Disconnect Air Break Switch Open Conductor TXF Secondary Breaker TXF Current Limiting Fuse

10

20

30

40

50

60

Percent of System SAIFI

Figure 6-14 Tree-Caused Customer Interruptions by Interrupting Device for One Northeastern Utility (Major Events Included)

Of tree-caused faults, a small number cause the greatest contribution to SAIFI and SAIDI. These are mainline faults. Restoration time is longest during major events and tree failures normally have the longest (and most expensive) cleanup and repair. Finch139 reported that on Niagara Mohawks system with over 2000 feeders in a given year, more than half of tree faults, half of customer outages, and half of customer hours of outage occurred on just 100 circuits (5% of feeders). Another utility in the northeast shows similar trends in Figure 5-26; half of the customer outage hours were from outages on just 4% of feeders. This data included major events, which tends to exaggerate the clustering effect.

6-20

Construction and Maintenance Programs to Reduce Tree Faults

100 80 60 40 20 0 0 20 40 60 80 100 10% of feeders contribute to 65% of treecaused SAIDI

Cumulative contribution to tree SAIDI

Cumulative percent of feeders

Figure 6-15 Allocation of Tree-Caused System SAIDI by Feeder for One Northeastern US Utility

Finch144 also reported that Niagara Mohawk found 3% of all tree-caused faults contributed to 52% of the total tree-caused SAIFI. Another northeastern US utility has similar results as shown in Figure 5-27.
Cumulative contribution to tree SAIDI
100 80 60 40 20 0 0 20 40 60 80 100

10% of tree faults contribute to 80% of treecaused SAIDI

Cumulative percent of treecaused faults

Figure 6-16 Allocation of Tree-Caused System SAIDI by Tree Outage for One Northeastern US Utility

Circuit voltage can also impact tree-caused faults. Tree-caused faults cause much less impact to customers on 5-kV class circuits. Finch reported that tree-cause outages on 2.4 to 4.16-kV
Finch, K., "Understanding Line Clearance and Tree Caused Outages," EEI Natural Resources Workshop, April 1, 2003a.
144

6-21

Construction and Maintenance Programs to Reduce Tree Faults

circuits averaged 79 customers per tree outage, but 7.6 to 13.2-kV circuits averaged 206 customers out per outage. Table 5-4 shows similar trends for another northeastern utility. Contrary to the widely held belief that 5-kV class circuits have much lower fault rates from trees, this utility had similar fault rates; the main difference is that faults impact less customers on the lower voltage circuits. Table 5-5 shows data from a southeastern utility that shows similar treecaused fault rates on 15- and 25-kV class systems.
Table 6-5 Tree Faults by Voltage Class for a Northeastern US Utility Tree SAIDI, minutes 4 22 34 Tree fault rate per 100 miles per year 11.6 9.6 10.8

Voltage 2.4 kV

4.16 kV 13.8 kV

Includes major storms

Table 6-6 Tree Faults by Voltage Class for a Southeastern US Utility Voltage class 15 kV 25 kV Tree fault rate per 100 miles per year With major storms 68 51 Without major storms 27 16

These findings strongly suggest that targeting should help improve power quality and more efficiently manage tree-maintenance budgets. Target tree maintenance and other treeimprovement strategies (like covered wire or increased spacings) to circuits with the most outages. To do that, focus on the following:

Mainline portion of circuits Circuits with more customers Circuits with a history of tree faults Circuits with higher voltage

6-22

Construction and Maintenance Programs to Reduce Tree Faults

Physics of Tree Faults Baltimore Gas and Electric145 pioneered some revealing tests on how trees cause faults. ECI146,147,148,149 and Florida Power Corporation150 followed this work with their own tests that showed much the same results. For a tree branch to cause a fault, the branch must bridge the gap between two conductors in close proximity, which usually must be sustained for more than one minute. A tree touching just one conductor will not fault at distribution voltages. The tree branch must cause a connection between two bare conductors (it can be phase to phase or phase to neutral). A tree branch into one phase conductor normally draws less than one amp of current under most conditions, this may burn some leaves, but it wont fault. On small wires in contact with a tree, the arcing to the tree may be enough to burn the wire down under the right conditions. While the tree in contact with one wire wont fault the circuit, there are some safety issues with trees in contact with overhead conductors. A fault across a tree branch between two conductors takes some time to develop. If a branch falls across two conductors, arcing occurs at each end where the wire is in contact with the branch. At this point in the process, the current is small (the tree branch is a relatively high impedance). The arcing burns the branch and creates carbon by oxidizing organic compounds. The carbon provides a good conducting path. Arcing then occurs from the carbon to the unburned portion of the branch. A carbon track develops at each end and moves inward. Once the carbon path is established completely across the branch, the fault is a low-impedance path (now the current is highit is effectively a bolted fault). It is also a permanent fault. If a circuit breaker or recloser is opened and then reclosed, the low-impedance carbon path will still be there unless the branch burns enough to fall off of the wires. Some other notable electrical effects include:

It makes little difference if the branch is wet or dry. Live branches are more likely to fault for a given voltage gradient, but dead branches are more likely to break and come in contact with the line. Little branches can burn through and fall off before the full carbon track develops. So, minor leaf and branch burning does not cause faults. The likelihood of a fault depends on the voltage gradient along the branch (see Figure 5-28).

Rees, W. T., Birx, T. C., Neal, D. L., Summerson, C. J., Tiburzi, F. L., and Thurber, J. A., "Priority Trimming to Improve Reliability," ISA conference, Halifax, NS, 1994. 146 Appelt, P. J. and Goodfellow, J. W., "Research on how trees cause interruptions - applications to vegetation management," Rural Electric Power Conference, 2004. 147 Goodfellow, J., "Understanding the Way Trees Cause Outages," 2000. http://www.eciconsulting.com/images/wnew/trees.pdf. 148 Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001. 149 Finch, K., "Understanding Line Clearance and Tree Caused Outages," EEI Natural Resources Workshop, April 1, 2003a. 150 Williams, C., "Tree Fault Testing on a 12 kV Distribution System," IEEE/PES Winter Power Meeting presentation to the Distribution Subcommittee, 1999.

145

6-23

Construction and Maintenance Programs to Reduce Tree Faults

The time it takes for a fault to occur also depends on the voltage gradient (see Figure 5-29). Lower voltage circuits are much more immune to flashovers from branches across conductors. A 4.8-kV circuit on a 10-foot crossarm has about a phase-to-phase voltage gradient of 1 kV/foot, very unlikely to fault from tree contact. A 12.47-kV circuit has a 2.7 kV/foot gradient, which is more likely to fault. Tests by ECI151 found that branch characteristics affected the probability of failure. Thicker branches were more likely to flash over, and live branches were also more likely to flash over for a given voltage gradient. ECI found no significant difference between naturally occurring growth and suckers (a secondary shoot produced from the base that often grows quickly). Moisture factor was less of a factor than one might guess. Surface moisture was less of a factor: it may make the fault occur more quickly but not make the fault more 152 likely. ECI did find differences between species. Florida Power Corp. found variation: in their tests palm limbs faulted fastest (one minute in their setup), and pine limbs lasted the longest (15 minutes).
100

Percentage of samples faulted

50

0 0 5 10

Voltage gradient across the branch, kV/ft

0 10 20 30

kV/m
Data source: Goodfellow, J., "Understanding the Way Trees Cause Outages," 2000. http://www.eci-consulting.com/images/wnew/trees.pdf.

Figure 6-17 Percentage of Samples Faulted Based on the Voltage Gradient Across the Tree Branch

Goodfellow, J., "Understanding the Way Trees Cause Outages," 2000. http://www.eciconsulting.com/images/wnew/trees.pdf. 152 Williams, C., "Tree Fault Testing on a 12 kV Distribution System," IEEE/PES Winter Power Meeting presentation to the Distribution Subcommittee, 1999.

151

6-24

Construction and Maintenance Programs to Reduce Tree Faults

150

Time of fault [seconds]

100

t = 1880/(kV/ft)3 = 53.2/(kV/m)3

50

0 0 5 10

Voltage gradient across the branch, kV/ft

0 10 20 30

kV/m
Data source: Goodfellow, J., "Understanding the Way Trees Cause Outages," 2000. http://www.eciconsulting.com/images/wnew/trees.pdf.with the curvefit added

Figure 6-18 Time to Fault Based on the Voltage Gradient Across the Tree Branch

These effects reveal some key issues:

Trimming around the conductors in areas with a heavy canopy does not prevent tree faults. Traditionally, crews trim a hole around the conductors with about a 10-ft (3-m) radius. If there is a heavy canopy of trees above the conductors, this trimming strategy performs poorly since most faults are caused by branches falling from above. Vertical construction may help since the likelihood of a phase-to-phase contact by falling branches is reduced. Candlestick or armless designs are more likely to flash over because of tighter conductor-toconductor spacings. Three-phase construction is more at risk than single-phase construction.

At transmission voltages, voltage gradients are high enough to cause a fault from a tree touching just one conductor. Hoffman et al.153tested a 220-kV circuit in contact with a poplar tree across 8 m (26 ft). Burning and arcing started at the tip of the tree and worked its way to a full short circuit at the bottom of the tree in 30 to 40 seconds. That is a voltage gradient of 4.9 kV/ft (15.9 kV/m), which is also in the range of Figure 5-29. Hoffman et al.154 also tested a 110-kV circuit, but arcing lasted twice as long before completely shorting out. They also tested a 20-kV circuit which did not flash when contacting one phase; after burning near the tip of the tree for 10 to 15 minutes, the small carbonized branches broke, and the tree became disconnected from the line. We are unaware of any tests on 34.5-kV circuits or other higher voltage distribution or lower voltage subtransmission lines. Given that a 34.5-kV circuit has 20 kV from line to ground, phase153

Hoffmann, E., Rheinbaben, H. v., and Stosser, B., "Trees in Electrical Contact with High Voltage Lines," CIGRE Session, 22-03, Sept. 1984. 154 Hoffmann, E., Rheinbaben, H. v., and Stosser, B., "Trees in Electrical Contact with High Voltage Lines," CIGRE Session, 22-03, Sept. 1984.

6-25

Construction and Maintenance Programs to Reduce Tree Faults

to-earth-level distances of at least 20 to 30 feet have a voltage gradient of less than one 1 kV/ft, a gradient unlikely to cause a fault according to the ECI tests (but extrapolation to longer distances might be invalid).

Covered Conductors and Other Construction Approaches to Reduce Tree Faults

This chapter considers various construction options to make circuits more resistant to tree-caused faults. This chapter will focus mainly on covered conductors and spacer cable, two insulated systems that are widely used to make overhead construction less susceptible to faults from trees. Underground Circuits An obvious approach to controlling tree-caused faults is to place T&D circuits underground. Underground circuits are almost immune to tree-caused faults (not completely though: uprooted trees can still disrupt underground facilities). Utilities save on maintenance costs as underground circuits do not need regular tree pruning. Also, storm restoration and repair costs are less, since most storm-related damage is from trees. Underground versus overhead construction has many tradeoffs in cost, power quality, safety, workability, and maintenance. Exposure to trees is just one of many variables that utilities should consider for the most appropriate type of circuit for a given location. Circuit Location To avoid tree falling faults, placing the line in locations not near trees or cutting down the large trees within falling distance of the line is possible but not very practical in many locations. Furthermore, it is noteworthy that in higher lightning locations placing lines out in the open could actually worsen the fault rate since the line is now exposed to more direct lightning strikes. Covered Conductors, Spacer Cable, and Aerial Cable Utilities with heavy tree cover often use covered conductors, conductors with a thin insulation covering (Figure 5-41 shows an example). The covering is not rated for full conductor line-toground voltage, but it is thick enough to reduce the chance of flashover when a tree branch falls between conductors. Covered conductor is also called tree wire or weatherproof wire. Tree wire also helps with animal faults and allows utilities to use armless or candlestick designs or other tight configurations. Tree wire is available with a variety of covering types. The insulation materials polyethylene, XLPE, and EPR are common. Insulation thicknesses typically range from 30 to 150 mils (1 mil = 0.001 in = 0.00254 cm); see Table 5-10 for typical thicknesses. From a design and operating viewpoint, covered conductors must be treated as bare conductors according to the National Electrical Safety Code (NESC)155, with the only difference that tighter conductor spacings are allowed. There are various grades of insulation used for the covering.
155

IEEE C2-2000, National Electrical Safety Code.

6-26

Construction and Maintenance Programs to Reduce Tree Faults Table 6-7 Typical Covering Thicknesses of Covered All-Aluminum Conductor

Diameter, inches Size AWG or kcmil 6 4 2 1 1/0 2/0 3/0 4/0 4/0 266.8 336.4 336.4 397.5 477 556.5 636 795 Cover Thickness in mils 30 30 45 45 60 60 60 60 60 60 60 80 80 80 80 95 95

Strands 7 7 7 7 7 7 7 7 19 19 19 19 19 37 37 61 61

Bare 0.178 0.225 0.283 0.318 0.357 0.401 0.451 0.506 0.512 0.575 0.646 0.646 0.702 0.771 0.833 0.891 0.997

Covered 0.238 0.285 0.373 0.408 0.477 0.521 0.571 0.626 0.632 0.695 0.766 0.806 0.862 0.931 0.993 1.081 1.187

Spacer cable and aerial cables are also alternatives that perform well in treed areas (Figure 5-42). Spacer cables are a bundled configuration using a messenger wire holding up three phase wires that use covered wire. Aerial cables have fully-rated insulation just like underground cables.

6-27

Construction and Maintenance Programs to Reduce Tree Faults

Courtesy of Duke Power Figure 6-19 Example of a Compact Armless Design Using Covered Conductors

Figure 6-20 Example of a Spacer Cable Run Through Trees

Other advantages of covered conductors include:

6-28

Construction and Maintenance Programs to Reduce Tree Faults

SpacingsThe NESC156 allows tighter conductor spacings on structures with covered conductors. Tighter spacings have aesthetic advantages. Also, more conductors can be placed in proximity, making it easier to build multiple-circuit lines, including underbuilt distribution. Spacer cables and aerial cables allow even more flexibility in squeezing more circuits on a pole structure. Animal-caused faultsCovered conductors add another line of defense against squirrels and other animals. Covering jumpers and other conductors that are near grounded equipment is the application that is most effective at reducing animal-caused faults. Fire reductionCovered conductors reduce the chances of fires starting from arcing between conductors and trees and other debris on the power line. Wildfire prevention is the main justification for using covered conductors in Australia157.

Safety is sometimes cited as a reason for using tree wire, but covered conductor systems do not necessarily offer safety advantages, and in some ways the covering is a disadvantage. Even 158 though Landinger et al. found small leakage currents through covered wires, they correctly point out that it doesnt cover all scenarios: covered conductors may reduce the chance of death from contact in some cases, but they are in no way a reliable barrier for protection to line workers or the public. Covered conductor circuits are more likely than bare-wire circuits to lead to downed-wire scenarios with a live distribution conductor on the ground. And if a covered wire does contact the ground, it is less likely to show visible signs that it is energized such as arcing or jumping which would help keep bystanders away. Additionally, with the use of covered conductors and spacer cables for preventing tree faults, preventing a fault is not always a good thing! If the weight of a tree deeply sags a covered conductor down to within reach of pedestrians, but because of its covering a fault does not occur, then the covered conductor may remain energized posing a public safety issue. On the other hand, with bare conductors if it is pulled down to this degree then a fault is more likely and an upstream protective device is likely to interrupt the circuit and de-energize the conductor posing less hazard to the public. The covering may also make a high-impedance fault less likely to transition to a low-impedance fault. If a downed phase conductor comes in contact (either intermittent or sustained) with a metallic object, the covering may prevent flashover for some time. Covered conductor systems have additional tradeoffs to be aware of. They are more susceptible to damage from fault arcs, they may cause radio frequency interference if the correct insulator tie is not used, and conductor corrosion is more likely. These topics will be discussed in more detail in a subsequent section.

IEEE C2-2002, National Electrical Safety Code. Barber, K., "Improvements in the Performance and Reliability of Covered Conductor Distribution Systems," International Covered Conductor Conference, Cheshire, UK, January, 1999. 158 Landinger, C. C., McAuliffe, J. W., Clapp, A. L., Dagenhart, J. B., and Thue, W. A., "Safety Considerations of Aerial Systems Using Insulated and Covered Wire and Cable," IEEE Transactions on Power Delivery, vol. 12, no. 2, pp. 1012-16, April 1997.
157

156

6-29

Construction and Maintenance Programs to Reduce Tree Faults

Industry Performance Data Good fault data is hard to find comparing fault rates of bare wire with covered wire. European experience with covered conductors suggests that covered-wire fault rates are about 75% less than bare-wire fault rates. In Finland, fault rates on bare lines are about 3 per 100 km/year on bare and 1 per 100 km/year on covered wire159. Table 5-11 shows data for Connecticut Light and Power comparing the performance of bare-wire construction to other constructions (this is also likely to be the source of the data cited by 160 Hendrix ). The bare-wire construction had much higher interruption rates than did the covered conductor, both tree-caused interruptions and other interruptions. Note that this data likely overstates the performance difference between bare and covered conductors. At CL&P, tree wire is used on new construction; the bare wire population is significantly older than the covered wire population. Most is 40+ year-old copper. Many of the additional failures on the bare wire system may have been age-related rather than being a function of the covering. CL&P documented several deficiencies on the older structures, including: excessive sag, rotted crossarms, and multiple splices. Also, most of their bare wire is on laterals rather than mainlines. This may influence the interruption rate comparisons, because fused laterals may have more interruptions from lightning and other temporary faults that could blow the tap fuse whereas a recloser or circuit breaker may successfully clear the same fault on the mainline.
Table 6-8 Interruption Rates for Various Constructions for CL&P Outages Per 100 Miles Per Year Bare Wire Trees Animals/Birds Lightning Equipment failure Unknown Others TOTALS 24.9 20.9 5.2 5.7 1.2 19.8 77.7 Tree Wire 3.6 3.0 1.0 0.7 0.8 2.6 11.6 Spacer Cable 1.1 2.4 1.6 1.8 0.8 2.7 10.5 Aerial Cable 6.0 2.4 3.7 10.6 2.4 8.2 33.3 Underground 0.3 0.9 1.2 13.1 0.9 5.5 21.8

Source: Connecticut Light and Power Company (CL&P) Transmission and Distribution Reliability Performance (TDRP) reports, submitted to the Connecticut Department of Public Utility Control Department, 2000 2004.

In South America, both covered wire and a form of aerial cable have been successfully used in treed areas161. The Brazilian company CEMIG found that spacer cable faults were lower than
Hart, B., "HV Overhead Line - the Scandanavian Experience," Power Engineering Journal, pp. 119-23, June 1994. 160 Hendrix, "Reliability of Overhead Distribution Circuits," August 1998. 161 Bernis, R. A. O. and de Minas Gerais, C. E., "CEMIG Addresses Urban Dilemma," Transmission and Distribution World, vol. 53, no. 3, pp. 56-61, March 2001.
159

6-30

Construction and Maintenance Programs to Reduce Tree Faults

bare-wire circuits by a 10 to 1 ratio (although the article didnt specify if this included both temporary and permanent faults). The aerial cable faults were lower than bare wire by a 20 to 1 ratio. The effect on interruption durations is shown in Table 5-12. Several spacer cables or aerial cables can be constructed on a pole. Spacer cables and aerial cables have some of the same burndown considerations as covered wire. Spacer cable construction does have a reputation for being hard to work with. Both spacer cable and aerial cable costs more than bare wire. CEMIG estimated that the initial investment was returned by the reduction in tree trimming. They did minimal trimming around aerial cable (an estimated factor of 12 reduction in maintenance costs) and only minor trimming around spacer cable (an estimated factor of 6 reduction in maintenance costs).
Table 6-9 Comparison of the Reliability Index SAIDI Construction Bare wire Spacer cable Aerial cable SAIDI, hours 9.9 4.7 3.0

SAIDI = average hours of interruption per customer per year Source: Bernis, R. A. O. and de Minas Gerais, C. E., "CEMIG Addresses Urban Dilemma," Transmission and Distribution World, vol. 53, no. 3, pp. 56-61, March 2001.

The utility data on types of tree faults can give us some idea of the maximum benefit from covered conductors. Depending on the utility, from a low of about 23% (Baltimore Gas and Electric) to a high of over 70% (Duke Power and BC Hydro) of tree faults were due to mechanical damage from large branches or entire trees falling on circuits. If we assume that the conductor covering will not affect mechanical damage from faults, then the best that a covered conductor will do is to reduce tree-caused faults by 30% (for utilities with a high percentage of mechanical damage) to 75% (for utilities with mainly growth or small limb contacts). This assumes application of covered conductors at the same spacings as bare conductors. If tighter spacings are used for covered conductors (often done), then the reduction in tree-caused faults may not be as great, but this is speculative as there is no industry data or testing to allow us to estimate the differences. A conductor covering may slightly increase the likelihood of mechanical damagethe covering increases the wires weight and mechanical load on the conductor, so it takes less force from a branch or tree to cause mechanical damage. Using the same reasoning, the extra ice loading on a covered conductor (due to increased surface area) could also increase the likelihood of damage from trees during ice storms. On the other hand, spacer cable systems can be more immune to mechanical damage from tree limbs. The combination of the high-strength messenger cable along with the tightly-bundled phase conductors is much stronger than a single conductor. With spacer cable, the force (weight) of the tree on the wires is more likely to be distributed amongst several wires (the neutral messenger and the three phases) than with crossarm, armless, or candlestick feeder designs. Furthermore, the numerous spacer insulator brackets may have more strength where they attach to the pole than standard pole top pin insulators. 6-31

Construction and Maintenance Programs to Reduce Tree Faults

Covered Wire Issues Arc Damage and Burndowns Fault-current arcs can damage overhead conductors, especially covered conductors. The arc itself generates tremendous heat, and where an arc attaches to a conductor, it can weaken or burn conductor strands. On distribution circuits, two problem areas stand out: 1. Covered conductorCovered conductor (also called tree wire or weatherproof wire) holds an arc stationary. Because the arc cannot move, burndowns happen faster than with bare conductors. 2. Small bare wire on the mainsSmall bare wire (less than 2/0) is also susceptible to wire burndowns, especially if laterals are not fused. Several utilities have had burndowns of covered conductor circuits when the instantaneous trip 162,163 . If a burndown on the main line occurs, all was not used or was improperly applied customers on the circuit will have a long interruption. In addition, it is a safety hazard. After the conductor breaks and falls to the ground, the substation breaker may reclose. After the reclosure, the conductor on the ground will probably not draw enough fault current to trip the station breaker again. This is a high-impedance fault that is difficult to detect. Covered conductor is susceptible to burndowns because when a fault current arc develops, the covering prevents the arc from moving. The heat from the arc is what causes the damage. Although ionized air is a fairly good conductor, it is not as good as the conductor itself, so the arc gets very hot. On bare conductors, the arc is free to move, and the magnetic forces from the fault cause the arc to move (in the direction away from the substationthis is called motoring). The covering constricts the arc to one location, so the heating and melting is concentrated on one part of the conductor. If the covering is stripped at the insulators and a fault arcs across an insulator, the arc motors until it reaches the covering, stops, and burns the conductor apart at the junction (see Figure 5-43 for an example of such damage). A party balloon, lightning, a tree branch, a squirrelany of these can initiate the arc that burns the conductor down.

Barker, P. P. and Short, T. A., "Findings of Recent Experiments Involving Natural and Triggered Lightning," IEEE/PES Transmission and Distribution Conference, Los Angeles, CA, 1996. 163 Short, T. A. and Ammon, R. A., "Instantaneous Trip Relay: Examining Its Role," Transmission and Distribution World, vol. 49, no. 2, 1997.

162

6-32

Construction and Maintenance Programs to Reduce Tree Faults

Courtesy of Duke Power Figure 6-21 Conductor Damage From Arcing Where the Conductor Cover Begins

Burndowns are most associated with lightning-caused faults, but its the fault current arc, not the lightning, that burns most of the conductor. Lightning triggers the arc; see Figure 5-22 for an example where a conductor was damaged by a puncture as the conductor flashed to the insulator tie. Lee et al.164 reported on a survey of 390 fallen covered conductor cases by the Pennsylvania Power and Light Company (PP&L). They found that over half of these cases were caused by lightning with trees as the second leading cause. Lightning and trees together accounted for 75% of burndown causes, and no other single cause exceeded 5%.

Courtesy of Duke Power Figure 6-22 Conductor Damage From Arcing

The conductor damage is a function of the duration of the fault and the current magnitude. Burndown damage from a fault arc occurs much more quickly than conductor annealing.

Lee, R. E., Fritz, D. E., Stiller, P. H., Kilar, L. A., and Shankle, D. F., "Prevention of Covered Conductor Burndown on Distribution Circuits," American Power Conference, 1980.

164

6-33

Construction and Maintenance Programs to Reduce Tree Faults

What we would like to do is plot the arc damage characteristic as a function of time and current along with the time-current characteristics of the protective device (whether it be a fuse or a recloser or a breaker). Doing this, we can check that the protective device will clear the fault before the conductor is damaged. Unfortunately, such arc damage data for different conductor sizes as a function of time and current is limited. Table 6-10 summarizes burndown characteristics of some bare and covered 165 conductors based on tests by Baltimore Gas & Electric . Figure 6-23 shows this same data on time-current plots along with a 100 K fuse total clearing characteristic. For conductor sizes not given, take the closest size given in Table 6-10, and scale the burndown time by the ratio of the given conductor area to the area of the desired conductor.

Goode, W. B. and Gaertner, G. H., "Burndown Tests and Their Effect on Distribution Design," EEI T&D Meeting, Clearwater, FL, Oct. 14-15, 1965.

165

6-34

Construction and Maintenance Programs to Reduce Tree Faults

Table 6-10 Burndown Characteristics of Various Conductors Duration, 60-Hz cycles Current, A
#6 Cu covered 100 200 360 1140 #4 Cu covered 200 360 1400 4900 #4 Cu bare 380 780 1300 4600 #2 ACSR covered 750 1400 4750 9800 #2 ACSR bare 1350 4800 9600 15750 1/0 Cu covered 480 1300 4800 9600 1/0 Cu bare 1400 4800 9600 15000 3/0 ACSR covered 1400 1900 3300 4800

Min
48.5 20.5 3.5 1.5 26.5 11 4.5 1 24.5 6 3.5 1 8 10 3.5 2 38 10 4.5 1 13.5 7 4 2 20.5 3.5 4 3 35 16 10 2

Max
55.5 24.5 4.5 1.5 36.5 12.5 5.5 1.5 32.5 9 7 1.5 9 9 4.5 2 39 11.5 5 1.5 20 15.5 5 2.5 29.5 7 6 4 38 17 12 3

Other
51 22 4.5 1.5 28 12 5.5 1.5 28.5 8 4.5 1 14 14 4 NA 40 10 6 NA 18 9 4.5 2.5 22.5 4.5 6 3.5 37 16.5 11 3

Curvefit
t=858/I1.51

t=56.4/I

0.92

t=641/I

1.25

t=15.3/I

0.65

t=6718/I

1.26

t=16.6/I

0.65

t=91/I

0.78

t=642600/I

1.92

Data source:. Goode, W. B. and Gaertner, G. H., "Burndown Tests and Their Effect on Distribution Design," EEI T&D Meeting, Clearwater, FL, Oct. 14-15, 1965.

6-35

Construction and Maintenance Programs to Reduce Tree Faults

Copper Bare 1.0 #4 0.1 1/0 #6

Copper Covered

#4

1/0

ACSR Bare 1.0

Time, seconds

ACSR Covered 4/0 3/0336.4 #2 3/0

#2

0.1

AAC Bare 1.0 500 350 0.1

ACAR Bare 4/0 336.4

0.1

10 0.1 Current, kA

10

The dashed line is the total clearing time for a 100 K fuse. Data source: Goode, W. B. and Gaertner, G. H., "Burndown Tests and Their Effect on Distribution Design," EEI T&D Meeting, Clearwater, FL, Oct. 14-15, 1965.

Figure 6-23 Burndown Characteristics of Various Conductors

Controlling Arc Damage The main ways to control arc damage on covered conductors and spacer cable is to limit the duration and magnitude of fault current. Options to manage fault currents include:

Transformer impedanceSpecifying a higher-impedance substation transformer limits the fault current. Normal transformer impedances are around 8%, but utilities can specify impedances as high as 20% to reduce fault currents. Split substation busMost distribution substations have an open tie between substation buses, mainly to reduce fault currents (by a factor of two).

6-36

Construction and Maintenance Programs to Reduce Tree Faults

Neutral reactorA reactor in the substation transformer neutral limits ground fault currents. Even though the neutral reactor provides no help for phase-to-phase or three-phase faults, it provides much of the benefits of other methods of fault reduction. Neutral reactors cost much less than line reactors (another option). Ground faults are the most common fault; and for many types of single-phase equipment, the phase-to-ground fault is the only possible failure mode. On the downside, a neutral reactor has a cost and uses substation space, and a neutral reactor reduces the effectiveness of the grounding system. Fault-current limitersSeveral advanced fault-current limiting devices have been 166 designed . Most use some sort of nonlinear elementsarresters, saturating reactors, superconducting elements, or power electronics such as a gate-turn-off thyristorto limit the fault current either through the physics of the device or through computer control. Since most distribution systems have managed fault currents sufficiently well, these devices have not found a market.

Overcurrent protective device selection, placements, and settings can impact the likelihood of damage on covered conductors and spacer cables. Consider the following options to more effectively use overcurrent devices to minimize arc damage:

Fuse savingUsing a fuse blowing scheme can increase burndowns because the fault duration is much longer on the time-delay relay elements than on the instantaneous element. With fuse saving, the instantaneous relay element trips the circuit faster and reduces conductor damage. For more information on fuse saving versus fuse blowing schemes, see EPRI 1001665167. Fuse ALL tapsLeaving smaller covered conductors unprotected is a sure way of burning down conductors. Tighter fusingNot all fuses protect some of the conductor sizes used on taps. Faster fuses reduce the chance of burndowns.

Finally, consider using bigger and/or stronger conductors to better withstand arcing. Bigger conductors take longer to burn down. Doubling the conductor cross-sectional area approximately doubles the time it takes to burn the conductor down. Using a stronger conductor such as ACSR may also help reduce the chance of downed conductors in that the steel messenger may still hold up the conductor even if a good portion of the aluminum is burned away. Arc Protective Devices Arc protective devices (APDs) are sacrificial masses of metal attached to the ends of where the covering is stripped (see Figure6-24 for an example). The arc end attaches to the mass of metal, which has a large enough volume to withstand much more arcing than the conductor itself. Figure 6-25 shows an example of an APD that has absorbed the energy from a fault-current arc; based on the loss of mass, this device could absorb several more such fault arcs before the

166

EPRI EL-6903, Study of Fault-Current-Limiting Techniques, Electric Power Research Institute, Palo Alto, California, 1990. 167 EPRI 1001665, Power Quality Improvement Methodology for Wires Companies, EPRI, Palo Alto, CA, 2003.

6-37

Construction and Maintenance Programs to Reduce Tree Faults

conductor was endangered. Lee et al.168reported work by PP&L to study burndowns and test arc protective devices as a solution. Based on this work, their standard practice is to strip the covering at each insulator and install APDs169. APDs only need to be installed on the load-side of the covering. The motoring action of the arc will push the arc away from the source, and it will attach to the bare conductor where the covering starts on the load side. Utilities can also specify installation of APDs on both sides of the stripped section. This is appropriate on circuits that can be operated as a loop, and it also eliminates the possibility of crew mistakes on which side is the load side. Due to low material cost, APDs are inexpensive when installed with the line, but they can be expensive to retrofit because of manpower issues. If a line crew is already set up on a structure, that is another cost-effective time to add APDs.

Courtesy of Duke Power

Figure 6-24 Arc Protective Devices

Lee, R. E., Fritz, D. E., Stiller, P. H., Kilar, L. A., and Shankle, D. F., "Prevention of Covered Conductor Burndown on Distribution Circuits," American Power Conference, 1980. 169 Lee, R. E., Fritz, D. E., Stiller, P. H., and Shankle, D. F., "Prevention of Covered Conductor Burndown on Distribution Circuits," Electrical World, pp. 107-9, December 1981.

168

6-38

Construction and Maintenance Programs to Reduce Tree Faults

Courtesy of Duke Power Figure 6-25 An Arc Protective Device That Has Operated

Lightning Protection and Covered Conductors A number of publications have advocated auxiliary lightning protection as a means of reducing the chance of covered conductor damage170. Even though most covered-conductor burndown scenarios may be initiated by lightning, it is difficult to stop a flashover from direct strikes. The most common approaches are with:

Surge arrestersSurge arresters can be used to protect overhead T&D circuits from flashover. For significant improvement from direct strikes, the arresters must be used at tight spacings, either at every pole or every other pole. Arresters at wider spacings may not limit direct-strike flashovers, but they will help reduce induced-voltage flashovers for lines with 171 low insulation levels. Current-limiting arcing hornsDeveloped in Japan, these are small-block metal-oxide arresters with an air gap made of an arcing horn172. The arcing horns are mainly for protection against induced voltages on lines with low insulation levels that are common in Europe and Japan (less than 100-kV critical flashover voltage). The arcing horn flashes over, and the metal-oxide element stops the power follow current. Since the metal-oxide blocks are so small, they cannot withstand direct strikes. They have not found use in North America.

Tapio, L., "Design of MV and HV Covered Conductor Overhead Lines," 17th International Conference on Electricity Distribution, CIRED, 2003. 171 IEEE Std. 1410-1997, IEEE Guide for Improving the Lightning Performance of Electric Power Overhead Distribution Lines. 172 Washino, M., Fukuyama, A., Kito, K., and Kata, K., "Development of current limiting arcing horn for prevention of lightning faults on distribution lines," IEEE Transactions on Power Delivery, vol. 3, no. 1, pp. 187-96, 1988.

170

6-39

Construction and Maintenance Programs to Reduce Tree Faults

Consider the tradeoffs carefully before considering additional lightning protection as a means of reducing covered conductor damage. First, if the lightning protection is done correctly, it may not offer much additional protection. Also, additional lightning protection is a significant expense and also introduces additional failure modes at each arrester location (from animals across the arrester bushing or from failure of the arrester itself). For more information on lightning protection, see IEEE Std. 1410-1997 and EPRI 1002188173. Wire Tie and Insulator Compatibility Pole structures with covered conductors can generate radio-frequency interference (RFI) if the insulator wire tie is not compatible with the covering. Power-line noise can be generated by conducting insulator ties separated by insulation from the line conductor. These scenarios include the following combinations:

Bare conductor tie on a covered line conductor thats not stripped at the insulator Insulated conductor tie on a bare or covered line conductor (see Figure 6-26)

Most power-line noise is from arcs, arcs across gaps on the order of 1 mm, usually at poor contacts. These arcs can occur between many metallic junctions on power-line equipment. Consider two metal objects in close proximity but not quite touching. The capacitive voltage divider between the conducting parts determines the voltage differences between them. The voltage difference between two metallic pieces can arc across a small gap. After arcing across, the gap can clear easily, and after the capacitive voltage builds back up it can spark over again. These sparkovers radiate radio-frequency noise. A conducting insulator tie in close proximity to the phase conductors creates a prime arcing scenario. A voltage can develop between the conducting insulator tie and the line conductor. The capacitance between the two is on the order of 30 to 50 pF, which is enough to charge the conducting tie relative to the line conductor174. The line covering may hold this voltage, but the covering may deteriorate or lightning may puncture it. Once the insulation has been bridged, repetitive arcing can occur across the air gap as the tie wire charges and then discharges into the line conductor. Arcing will further deteriorate the conductor insulation, possibly causing more 175 arcing. Vincent and Munsch also document a second cause of RFI from incompatible insulator ties: if the insulation deteriorates enough so that the conducting insulating tie touches the line conductor, then an insulating oxide layer can build between the two, leading to microsparking noise from breakdowns across this small gap.

EPRI 1002188, Power Quality Implications of Distribution Construction, Electric Power Research Institute, Palo Alto, CA, 2004. 174 Vincent, W. R. and Munsch, G. F., Power-Line Noise Mitigation Handbook, United States Navy Naval Graduate School and ARRL, 2002. 175 Vincent, W. R. and Munsch, G. F., Power-Line Noise Mitigation Handbook, United States Navy Naval Graduate School and ARRL, 2002.

173

6-40

Construction and Maintenance Programs to Reduce Tree Faults

For more detail on power-line RFI including finding sources and solutions, see Loftness176, NRECA 90-30177, and the US Navy/ARRL Power-Line Noise Mitigation Handbook178.

Source: Vincent and Munsch178 Figure 6-26 Example of a Covered Wire Tie on a Covered Conductor

The main problem with these partial discharges is that they cause radio interference. There has been speculation that these discharges could damage the conductor, but in tests by the Pennsylvania Power and Light Company (PP&L), Lee et al.179reported that in tests of different wire tie and insulator combinations, no evidence of conductor damage was found. To reduce radio interference with covered-conductor systems, use insulator ties that are compatible with the insulator:

Either strip the conductor at each insulator and use bare metallic insulator ties; or Leave the conductor covering on, and use nonconducting insulator ties.

For covered-conductors with conducting insulator ties, a retrofit is possible by stripping the insulation on one side and bonding the insulator tie to the conductor.

Loftness, M. O., AC Power Interference Field Manual, Percival Publishing, Tumwater, WA, 1996. NRECA 90-30, Power Line Interference: A Practical Handbook, National Rural Electric Cooperative Association, 1992. 178 Vincent, W. R. and Munsch, G. F., Power-Line Noise Mitigation Handbook, United States Navy Naval Graduate School and ARRL, 2002. 179 Lee, R. E., Fritz, D. E., Stiller, P. H., Kilar, L. A., and Shankle, D. F., "Prevention of Covered Conductor Burndown on Distribution Circuits," American Power Conference, 1980.
177

176

6-41

Construction and Maintenance Programs to Reduce Tree Faults

Some utilities argue that lines have better lightning protection if the covering is left on the conductor. While the improvement is marginal, there is some difference between different 180 covering and insulator tie combinations. Tests at Clarkson University of 15-kV class pin insulators in the 1980s found that keeping the cover on raises the critical flashover voltage from about 115 kV with bare wire to about 145 kV with the cover on using a preformed plastic tie (with a semiconductive tie or a polyethylene covered aluminum tie, the values were slightly less than this). For a direct strike, these differences should not matter, but for a weakly insulated line (with little wood or fiberglass), the extra insulation could help reduce induced-voltage flashovers, but for most North American designs, the difference in overall insulation is small. Direct strikes will still cause flashovers and possible damage; the most likely flashover point is where the insulation is weakest: at the insulator where the tie comes in contact with the covering (Figure 6-27). The covering will not add significant insulation to a structure with an insulator and a foot or more of wood or fiberglass.

Courtesy of Duke Power Figure 6-27 Example of Damage on a Covered Conductor From Flashover at the Insulator Tie

Other Covered-Conductor Issues Covered conductors are heavier, have a larger diameter, and have a lower strength rating. Relative to the same size bare conductor, a 477-kcmil all-aluminum conductor with an 80-mil XLPE conductor covering weighs 20% more, has a 17% larger outside diameter, and has a 10% lower strength rating. The ice and wind loading of a covered conductor is also higher than a comparable bare conductor. Both increase with increasing diameter. In the example comparing a 477-kcmil allaluminum conductor with an 80-mil XLPE covering, the loadings for the covered conductor versus a bare conductor increase as follows:
Baker, R. J., "Impulse Breakdown Characteristics of 13.2 kV Polyethylene Covered Conductor Distribution Systems," Master of Science Thesis, Clarkson University, 1984.
180

6-42

Construction and Maintenance Programs to Reduce Tree Faults

VerticalLoading due to ice and conductor weight increases 14%. HorizontalLoading due to wind increases 8%. ResultantLoading due the vertical and horizontal component increases 11%.

Another issue with covered conductors is the integrity of the covering. The covering may be susceptible to degradation due to ultraviolet radiation, tracking and erosion, and abrasion from rubbing against trees or other objects. Early covering materials, including the widely used PVC, were especially susceptible to degradation from ultraviolet light, from tracking, and from abrasion. Modern EPR or XLPE coverings are much less susceptible to degradation and should be much more reliable. Covered conductors are more susceptible to corrosion, primarily from water. If water penetrates the covering, it settles at the low points and causes corrosion (the water cant evaporate). On bare conductors, corrosion is rarerain washes bare conductors periodically, and evaporation takes take of moisture. Australian experience has found that complete corrosion can occur with 181 covered wires in 15 to 20 years of operation . Water enters the conductor at pinholes caused by lightning strikes, cover damage caused by abrasion or erosion, and at holes pierced by connectors. Temperature changes then cause water to be pumped into the conductor. Because of corrosion concerns, water-blocked conductors are better. Covered conductors have ampacities that are close to bare-conductor ampacities for the same operating temperature. Covered conductors are darker, so they absorb more heat from the sun but radiate heat better. The most significant difference is that covered conductors have less ability to withstand higher temperaturesthe insulation degrades. Polyethylene is especially prone to damage, so it should not be operated above 75C. EPR and XLPE may be operated up to 90C. A bare conductor may have a rating of as high as 100C. Mechanical Coordination of Construction Trees causing mechanical damage make up a large portion of tree-caused faults, and these are the faults that require the most time and expense to repair. One approach to reducing these faults is to target and remove hazard treesthose most likely to fall on T&D infrastructure. Another approach to reducing the impact of the damage is to coordinate the mechanical design such that when tree and large limb failures occur, equipment fails in a manner that is easier for crews to repair. When a tree falls on a line, crews will have an easier repair if it just breaks the conductors rather than breaking poles and other supports. The fault still occurs, but crews are able to more quickly repair the damage and restore service. Figure 6-28 shows an example of a hard-to-repair failure; if the conductors or insulator ties had broken first, the poles may have been left standing, and crews would have been able to repair it more quickly.

Barber, K., "Improvements in the Performance and Reliability of Covered Conductor Distribution Systems," International Covered Conductor Conference, Cheshire, UK, January, 1999.

181

6-43

Construction and Maintenance Programs to Reduce Tree Faults

Figure 6-28 A Pole Broken in Half by a Tree Falling Onto the Line Structure During a Windstorm

Spacer cable systems are an example that can cause a mechanical miscoordination. Spacer cable systems are quite strong, so they withstand some tree branch contact that an open-wire system would not. But, the spacer cable is strong enough that the conductors are less likely to be the weakest link. When a heavy tree does fall on the line, the spacer cable can break several poles, leading to a much longer repair time. BC Hydro182 has developed an approach to overhead structure design that considers the order of failure of equipment. Yu et al.183 developed methodologies for calculating conductor tensions under the stress of a large concentrated load (the falling tree or branch). With the tension information, they use a probabilistic approach to determining failures of components. Each components probability of failure is used to rank the likelihood of failure of each component. Then, once the weakest member is determined to fail, the stresses and probabilities are recalculated for the remaining components to determine what might fail next. This provides a sequence of failures for a given design. BC Hydro used this analytical approach to analyze several of their standard designs. They found the following general results:

Neither pole species, pole length, or pole classes affected results. Trees falling near midspan and those falling near a pole were similar. For tangent structures, with #2 ACSR, the phase and neutral failed first when a tree fell on either conductor. For 336.4-kcmil ACSR, the pole tended to fail first.

Kaempffer, F. and Wong, P. S., "Design Modifications Lessen Outage Threat," Transmission & Distribution World, October 1, 1996. 183 Yu, P., Wong, P. S., and Kaempffer, F., "Tension of Conductor Under Concentrated Loads," Journal of Applied Mechanics, vol. 62, pp. 802-9, September 1995.

182

6-44

Construction and Maintenance Programs to Reduce Tree Faults

For angle structures, the guy grip for the phase and the tie wire for the neutral usually failed first. For deadend structures, the guy grip tended to fail first.

Although this method of distribution design is not widely used, mechanical coordination should be given consideration in designing T&D structures to make them easier to repair during storms. Choice of conductor (AAC versus ACSR) also plays a roll. In some applications, ACSR may be strong enough to move the weakest link to a harder-to-repair supporting structure.

Utility Programs to Reduce Tree-Related Power Quality Problems

Tree Maintenance When trees contact utility equipment, damage is often extensive, and repair is expensive and time-consuming. Most of these vegetation faults are on distribution circuits, but transmission circuits are not immune as highlighted by the famous tree-caused fault on August 14, 2003 on one of FirstEnergy's 345-kV lines in Ohio. In addition to long-duration interruptions, the faults from trees cause voltage sags and can cause momentary interruptions. For most utilities, vegetation management is by far the largest maintenance item in the budget. So, in addition to improving power quality, more efficient tree maintenance and more treeresistant designs can also reduce maintenance. General Guidelines Vegetation management programs can be implemented more efficiently by more focused programs:

RemovalThis is the most effective fault-prevention strategy, and many homeowners are willing to have trees removed, especially dead or decaying trees. Danger treesTrimming/removal is most effective if trees and branches that are likely to fail are removed or trimmed to safe distances. This does take some expertise by tree trimming crews. Target circuitsAs with any fault-reduction program, efforts are best spent on the poorest performing circuits that affect the most customers. Target mainlinesThe most significant power quality impacts on customers are for tree faults on the three-phase mains, so prioritize tree maintenance to target those faults.

Consider the type of tree exposure affecting a line. Compare the circuits in Figure 6-29 that have significant tree overhangs to the circuits in Figure 6-30 that have exposure to growth. Which are worse? As shown by the data from several utilities, branch failure and tree failure are much more of a problem than tree growth. So, tree maintenance strategies should concentrate on removing overhangs and removing trees in danger of falling on circuits.

6-45

Construction and Maintenance Programs to Reduce Tree Faults

In general, more aggressive cutting strategies will have the most significant power quality and reliability impact. Tree removal can also be less costly than selective pruning, especially when the costs are evaluated over the life of the tree.

Figure 6-29 Circuits With Significant Tree Overhang

6-46

Construction and Maintenance Programs to Reduce Tree Faults

Figure 6-30 Circuits With Impending Tree Growth Contact

Vines are a special situation that requires special attention (Figure 6-31). Vines can grow very quickly, and they cause some of the most repeatable faults. Because of the repeatability, when responding to a vine-caused interruption, crews should exterminate the offending vines as best as possible (or flag the pole for later attention by vegetation crews). Repeatability is also a reason to have a separate outage cause code for vinesthis makes it easier to follow vine-caused faults for vegetation crews. Especially on circuit mainlines, vines should be targeted for removal because they are so fast growing and will cause repeated outages if not kept under control.

6-47

Construction and Maintenance Programs to Reduce Tree Faults

Figure 6-31 Circuits With Vines

Acceptable tree trimming (that is also still effective) is a public relations effort. Some strategies that help along these lines include:

Talk to residents prior to/during tree trimming. Get permission for removal of hazard trees outside of the normal trim zone. Trim trees during the winter. The community will not notice tree trimming as much when the leaves are not on the trees. Trim trees during storm cleanups. Right after outages, residents are more willing to accept tree maintenance. Clean up after trees are trimmed/removed. Offer free firewood.

Program Cost Tree trimming is expensivean EPRI survey found that utilities spend an average of about $10 per customer each year on tree trimming184 and Figure 6-32). Trimming can also irritate communities. It is always a dilemma that people dont want their trees trimmed, but they also dont want interruptions and other power quality disturbances.
184

EPRI TR-109178, Distribution Cost Structure -- Methodology and Generic Data, Electric Power Research Institute, Palo Alto, California, 1998.

6-48

Construction and Maintenance Programs to Reduce Tree Faults

$ pe r Cus tom e r $16 $11

$ pe r OH Circuit m ile $1,000

$610 $425
$4

High

Avg

Low

High

Avg

Low

$ pe r s um total of Cus tom e r M inute Inte rruption $0.28 $0.20 $0.15

$ pe r m illion k Wh $500 $400

$90

High

Avg

Low

High

A vg

Low

Data source: EPRI TR-109178, Distribution Cost Structure -- Methodology and Generic Data, Electric Power Research Institute, Palo Alto, California, 1998.

Figure 6-32 Utility Vegetation Management Costs of Five Utilities Surveyed

Figure 6-33 shows costs for other utilities from publicly available data (mainly regulatory filings). Costs vary significantly from utility to utility and reflect differences in tree coverage, load density (urban and suburban trimming is more difficult than rural tree maintenance), vegetation management cycle, and tree growth rates.

6-49

Construction and Maintenance Programs to Reduce Tree Faults

Vegetation Management Expenditure in US$/OH miles/year

Detroit Edison

$1,428

Commonwealth Edison

$1,145

Central Illinois Light Company

$845

Illinois Power

$686

Energex

$618

Dominion Virginia Power

$551

United Illuminating

$442
0 200 400 600 800 1000 1200 1400 160

Data sources: various public regulatory filings to state agencies Figure 6-33 Costs of Vegetation Management From Several Utilities

Figure 6-34 shows additional data on the costs of vegetation management programs and also ties that to performance using the SAIDI index. There is little direct correlation between spending and SAIDI between utilities. This is not surprising given wide variances in tree coverage, load and customer densities, and weather between utilities.

6-50

Construction and Maintenance Programs to Reduce Tree Faults

1.0
Portion exceeding the xaxis value

1.0

0.8 0.6 0.4 0.2 0.0 0 500 1000 1500 2000 2500 Vegetation maintenance expenses US$ per overhead mile per year

Portion exceeding the xaxis value

0.8 0.6 0.4 0.2 0.0 0 50 100 150

Treerelated SAIDI, minutes per year

Vegetation maintenance expenses US$ per overhead mile per year

2500 2000 1500 1000 500 0 0 50 100 150

Treerelated SAIDI, minutes per year

Data source: 2003 PA Consulting Benchmarking survey 185 Figure 6-34 Vegetation Management Costs Versus Performance

Tree-Trimming Cycle Choosing a tree-trimming cycle is tricky. Many utilities use a three to five-year cycle. Longer tree-trimming cycles should lead to higher fault rates. The optimal trimming cycle depends on:

Type of trees, growth rates, and growing conditions Community tolerance for trimming Economic assumptions, especially the chosen time value of money

BC Hydro, "Revenue Requirement Application 2004/05 and 2005/06, Chapter 7. Electricity Distribution and Non-Integrated Areas," 2003.

185

6-51

Construction and Maintenance Programs to Reduce Tree Faults

The following graphs characterizes one utilitys tree-caused outages from 1999 to 2003, including major storms. Each graph shows the effect of the time since the last tree trimming. So, the first datapoint in each graph is the given reliability index of all circuits that had tree maintenance during the previous year. One would expect that tree-caused outages would decrease immediately following tree maintenance, but the data does not show this. There is no strong trend for any of the benchmarks shown in the graphs.
0.25
SAIDI from tree contact, minutes SAIFI from tree contact

0.20 0.15 0.10 0.05 0.00 1 2 3 4

60

40

20

0 1 2 3 4

Years since last tree maintenance

Years since last tree maintenance

Figure 6-35 Tree Maintenance Effect on SAIFI and SAIDI for a Northeastern Utility

Tree outages/100 OH miles

15

10

0 1 2 3 4

Years since last tree maintenance

Figure 6-36 Tree Maintenance Effect on Tree Outage Rate Per Mile

This data is not unique. The following data is from another utility that also does not show a trend with regards to time since the last tree maintenance. One difference in these charts is that the numbers are for all faults, not just those caused by trees.

6-52

Construction and Maintenance Programs to Reduce Tree Faults

0.29 1.1

Faults per mile per year

1 2 3 4 5 6 7 8

0.28 0.27 0.26 0.25 0.24

SAIFI

1.0

0.9

0.8 1 2 3 4 5 6 7 8

Years since last tree trimming

Years since last tree trimming

Figure 6-37 Tree Maintenance Effect on Tree Outage Rate Per Mile for a Southeastern Utility

Correlating the effect of tree maintenance and performance can be tricky. Some effects that can interfere with such correlations include:

TargetingTargeting poorly performing circuits for maintenance can help improve customer power quality, but it makes it difficult to gauge the effects of maintenance programs. Maintenance approachSome utilities schedule vegetation maintenance using map sections, not by circuit, so it is impossible to correlate circuits to their performance. Budget and tree maintenanceVegetation management budgets vary considerably and so do pruning specifications. Both can impact different years differently. ReconfigurationsCircuit reconfigurations can make it difficult to reliably judge the history of tree fault impacts.

The main point of this data is that tree-caused outages do not increase dramatically with longer times between tree maintenance when analyzed over reasonable time periods. Quoting 186 Guggenmoos : Only a trim program that is substantially behind cycle results in increased outages. Might say that cycle trimming is not for reliability but public safety and the avoidance of higher costs associated with heavily pruning systems that have grown into conductors. Effect on faults and interruptions is not the only reason for selecting a maintenance cycle. Several other factors include:

Shock hazard Trees can (rarely) cause a possibly dangerous shocking hazard. A tree in contact with an energized phase conductor may create a touch potential hazard near the ground. For a tree in contact with one conductor, the resistance of a tree is high enough to remain a high-impedance connectionit will not draw enough current to operate a fuse or other protective device, but it may be enough to create significant step potentials. St. Clair187 reported two electrocutions from tree contacts to 12.47-kV lines, one a tree trimmer and

Guggenmoos, S., "Effects of Tree Mortality on Power Line Security," EEI Natural Resources Workshop, April 2003b. In the notes to the presentation. 187 St. Clair, J. G., 1999. Discussion to Daily [1999].

186

6-53

Construction and Maintenance Programs to Reduce Tree Faults

another at ground level where a man leaned against a tree that had fallen into conductors. The probability of developing a shock hazard is a function of the tree resistivity, the earth resistivity, and whether the tree is wet (see Daily188 and Short189). The most dangerous conditions are when the ground is wet and the tree is wet. Low earth resistivity grounds the tree better, so it draws more current from the line. Foot-to-earth contact resistances are also lower with drenched soil, which draws more current through your body. A wet treedue to rain or humid weatheris more dangerous as the trees resistance drops appreciably. Cutting trees more aggressively or more frequently helps reduce the shock probability.

Fire hazard For areas in high fire-danger areas, tree clearance requirements may be more severe, requiring more frequent maintenance. The State of California Rules for Overhead Electric Line Construction specifies at least an 18-inch spacing between conductors and vegetation for all distribution and transmission circuits. In addition, California (Public Resource Code Section 4293) requires the following clearances for circuits in any mountainous land, or in forest-covered land, brush-covered land, or grass-covered land:

For any line which is operating at 2,400 or more volts, but less than 72,000 volts, four feet. For any line which is operating at 72,000 or more volts, but less than 110,000 volts, six feet. For any line which is operating at 110,000 or more volts, 10 feet. To meet such spacing requirements normally requires more frequent vegetation maintenance.

Cost Longer maintenance cycles may actually cost more as the catch-up phase can be more expensive than maintaining a consistent budget. In a survey of three utilities, Environmental Consultants, Inc. (ECI) found that extending tree maintenance cycles beyond the optimum can increase overall costs190,191. If cycles are increased, costs are higher because (1) it takes more time for crews to prune when trees are in close proximity to conductors, (2) crews must do more hot-spot maintenance in response to trouble calls, and (3) crews have more mass of debris to clear and dispose of. ECI estimated that for each one dollar saved by extending maintenance cycles would require from $1.16 to $1.23 in spending if the cycle was extended one year past its optimum, and if the circuit is four years past the optimum, the catch-up cost is 1.47 to 1.69 times the cost originally saved. Storm repair Trees cause considerable damage during storms. Duke Power discovered that during the ice storm of 2002 that impacted their service territory, the circuits that had not been maintained in 13 years had five times the damage of circuits that had been maintained from 1 to 6 years ago192. Because of this, Duke justified increasing the vegetation management budget based on reducing storm repair costs.

Daily, W. K., "Engineering Justification for Tree Trimming [Power System Maintenance]," IEEE Transactions on Power Delivery, vol. 14, no. 4, pp. 1511-8, October 1999. 189 Short, T. A., Electric Power Distribution Handbook, CRC Press, Boca Raton, FL, 2004. 190 Browning, D. M. and Wiant, H. V., "The Economic Impacts of Deferring Electric Utility Tree Maintenance," Journal of Arboriculture, vol. 23, no. 3, May 1997. 191 Massey, B., "Deferring Tree Trimming: A Million-Dollar Mistake," Electrical World, pp. 56-7, September 1998. 192 Taylor, L., 2004b. Personal communication.

188

6-54

Construction and Maintenance Programs to Reduce Tree Faults

Politics Regulatory bodies are paying more attention to performance and vegetation management. Tree-maintenance cycle is an easy indicator for regulators to understand. If a utility decreases budgets and/or increases tree maintenance cycles and that coincides with decreased reliability or customer satisfaction, regulators may impose fines or mandate changes. For example, CILCO failed to meet its target tree maintenance cycle of four years (Table 6-11) and was severely reprimanded by the regulator in a Staff Report to the Illinois Commerce Commission on October 17, 2000; their tree-maintenance cycle had slipped to the equivalent of a ten-year cycle.
Table 6-11 Illinois Utility Tree Maintenance Cycle Targets Tree-maintenance target in years Utility Urban 4 3 3 4 4 3 to 4 3 Rural 5 4+ 4+ 4 4 3 to 4 3

Alliant-Interstate Power Company AmerenCIPS AmerenUE CILCO ComEd Illinois Power Company MidAmerican Energy Company

Source: Staff Report to the Illinois Commerce Commission on October 17, 2000

Hazard-Tree Programs Danger-tree or hazard-tree programs target those trees that are the largest threats to utility circuits. Tree trimming within a zone (+/- 10 feet for example) targets tree growth, but most tree outages are from trees or branches from outside of typical utility trim zones. Danger-tree programs target dead trees or trees with significant defects, even if they are out of the normal trim zone or right-of-way. Figure 5-54 shows some of the tree defects that led to tree faults in a study by the Niagara Mohawk Power Corporation. Dead trees are the most obvious candidates for hazard-tree removals. In a sample of permanent tree faults, Niagara Mohawk found 36% were from dead trees; and in another sample, Duke Power found 45% were from dead trees193.

Taylor, L., "The Illusion of Knowledge," Southeastern Electric Exchange Power Quality and Reliability Committee, Dallas, TX, 2003.

193

6-55

Construction and Maintenance Programs to Reduce Tree Faults

Decay, Rotted, Punky Codominant Stems or Leads Dead Along Side or Overhead Overhead and Overhanging (species) Cracks and Splits Open/Visible

10

15

20

Percent of defects causing permanent tree faults

Data source: Finch194 Figure 6-38 Defects Causing Tree Failure for the Niagara Mohawk Power Corporation

Targeting danger trees is highly beneficial, but requires expertise. In a careful examination of several cases where broken branches or trees damaged the system, 64% of the trees were 194 living . Finch also advises examining trees from the backside, inside the tree line (defects on that side are more likely to fail the tree into the line). Finch describes several defects that help signal danger trees (see Figure 6-38). Dead trees or large splits are easy to spot. Cankers (a fungal disease) or codominant stems (two stems, neither of which dominates, each stem at a branching point is approximately the same size) require more training and experience to detect. For identifying hazard trees, it also helps to know the types of trees that are prone to interruptionsthis varies by area and types of trees. Finch195 showed how Niagara Mohawk evaluated a sample set of tree outages in a study in 2000. Niagara Mohawk compared the tree species that caused faults to the tree species in New York state. They found that Black Locusts and Aspens are particularly troublesome; large, old roadside maples also caused more than their share of damage (see Table 6-12). Finch also reported much of the extra impact of Aspens on outages was due to hypoxilon canker, which their crews often overlooked as a defect. The sugarmaple faults were mainly from large, old roadside maples in serious decline.

194 195

Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001. Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001.

6-56

Construction and Maintenance Programs to Reduce Tree Faults

Table 6-12 Comparison of Trees Causing Permanent Faults With the Tree Population Species Ash Aspen Black Locust Black Walnut Red Maple Silver Maple Sugar Maple White Pine Source: Finch
196

Percent of outages 8 9 11 5 14 5 20 6

Percent of New York state population 7.9 0.6 0.3 N/A 14.7 0.2 12.0 3.3

In an informal survey of seven utilities, Guggenmoos197 found that most utility hazard-tree programs removed about five trees per mile of circuit, with the most intense programs removing 10 to 15 trees per mile. Note that while danger-tree programs can improve power quality, they are not a panacea. Tree outages will still occur regularly. Many tree faults are from weather that causes tree failures of otherwise healthy trees. Danger-tree programs must be ongoing programs. As Guggenmoos198 shows in detail, with tree mortality rates on the order of 0.5 to 3% annually and the sheer number of trees within striking distance of T&D circuits, a danger-tree program cannot be a one-time expenditure. Identifying Tree Defects While most hazard-tree programs are best directed by a professional forester, it is beneficial for anyone involved in distribution power quality field investigations to have some background knowledge of common tree defects. Fortunately, many common tree defects are relatively easy to identify. Although making predictions of future tree behavior for targeting hazard tree removal may require a trained forester, some basic background on common tree failures is often sufficient for the engineer investigating a power quality problem. Defects can often be linked to previous wounding, infestation, or undesirable growing conditions and are a visible sign that a tree has a disposition to fail. Just like any other structure, trees fail whenever the loading on them exceeds their mechanical strength. Defective trees will fail before
Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001. Guggenmoos, S., "Effects of Tree Mortality on Power Line Security," Journal of Arboriculture, vol. 29, no. 4, pp. 181-96, July 2003a. 198 Guggenmoos, S., "Effects of Tree Mortality on Power Line Security," Journal of Arboriculture, vol. 29, no. 4, pp. 181-96, July 2003a.
197 196

6-57

Construction and Maintenance Programs to Reduce Tree Faults

healthy trees because the defect lowers the mechanical strength of the surrounding wood thus weakening the tree. The most common tree defects are as follows:

Deadwood as the name implies, this is wood that has died. Deadwood is dry and brittle and cannot bend under load (wind, ice, etc.) and is therefore prone to breakage. This defect can range from individual branches to whole trees. Deadwood is often indicated by limbs or trees that do not have green, new growth leaves during the summer season such as that shown in Figure 6-39. Dangling dead braches are particularly dangerous.

Figure 6-39 Example of Deadwood

Cracks A crack is a deep split that extends through the bark and into the wood. Cracks are a sign that the tree has already started to fail. There are four types of cracks (Figure 6-40):

Shear Crack - A shear crack, as shown in Figure 6-40, separates the stem into two halves and carries a high risk of failure. Inrolled Crack The edges of this vertical crack are inrolled encompassing the bark and wood. Trees with inrolled cracks almost always suffer serious decay at the crack site. Ribbed Crack A ribbed crack is a fissure in a raised rib of wood along the length of the stem. Horizontal Crack These cracks run across the wood grain and are rare to find since they develop just before the tree fails.

6-58

Construction and Maintenance Programs to Reduce Tree Faults

Shear cracks

Inrolled crack
Figure 6-40 Examples of Cracks

Ribbed crack

Horizontal crack

Weak Branch Unions A weak branch union is a defect or weakness at the point at which two branches separate. A strong branch union is characterized by a small bark ridge line in the center of the union as shown in Figure 6-41. This indicates that the annual rings of the branch and stem are growing together creating a strong union. Weak unions lack this ridge

6-59

Construction and Maintenance Programs to Reduce Tree Faults

line indicating that bark may be growing into the union (Figure 6-42) or that the union is an offshoot from a previously damaged spot.

Notice the small ridge of bark in the center of the union Figure 6-41 Example of a Strong Branch Union

Notice the absence of a bark ridge line in the center of the union and the ingrown bark Figure 6-42 Example of a Weak Branch Union

Decay Decayed wood is the result of long-term exposure to decay-causing fungi. Decay is primarily an internal process that offers few outward indications and it can occur in the roots, stem, or branches. Decayed wood is always weaker than healthy wood and is therefore prone to failure. When present, outward signs of decay include discoloration, holes, and fungal fruiting bodies as shown in Figure 6-43.

6-60

Construction and Maintenance Programs to Reduce Tree Faults

Figure 6-43 Examples of Outward Signs of Tree Decay Discoloration, Holes, and Fungal Activity

Cankers A canker is a dead area of bark and/or cambium which stops a new annual ring from being added each year at that location. Since cankers do not allow tree growth they reduce the trees strength by limiting the amount of wood at that location. Several examples of cankers are shown in Figure 6-44.

Figure 6-44 Examples of Cankers

Root Problems Root damage or inadequate root anchoring results in the tree tipping over because it cannot anchor itself into the ground. Root problems can be caused by tree growth in a confined earth area (such as a city sidewalk), excavation or paving near the tree, fungal infection, drought, or flood. Some indications of root problems include leaning trees, reduced tree crown, exposed roots, and recent construction work near the tree(s) as shown in Figure 6-45, Figure 6-46, and Figure 6-47.

6-61

Construction and Maintenance Programs to Reduce Tree Faults

Figure 6-45 Example of Root Damage due to Excavation

Figure 6-46 Example of Crown Decline due to Root Damage

Figure 6-47 Examples of Leaning Trees due to Root Damage

Poor Tree Architecture Poor architecture is a growth pattern, such as that shown in Figure 6-48, that indicates structural imbalance or weakness in the tree. Poor tree

6-62

Construction and Maintenance Programs to Reduce Tree Faults

architecture develops over many years due to damage and poor environmental conditions. Trees that are leaning or have large branches that are out of proportion to the rest of the crown are particularly prone to failure.

Figure 6-48 Example of Poor Tree Architecture

Much of the above information came from the two following sources, both of which are recommended for further reading:

How to Recognize Hazardous Trees, USDA Forrest Service Northeastern Area, Report NA-FR-01-96. At the time of publishing, this report is available on the web at: http://www.na.fs.fed.us/spfo/pubs/howtos/ht_haz/ht_haz.htm#what Urban Tree Risk Management: A Community Guide to Program Design and Implementation, USDA Forrest Service Northeastern Area, Report NA-TP-03-03. At the time of publishing, this report is available on the web at: http://www.na.fs.fed.us/spfo/pubs/uf/utrmm

Further information can also be found in:

Fazio, J., How to Recognize and Prevent Hazard Trees. Tree City USA Bulletin No. 15. Nebraska City, NE: National Arbor Day Foundation 1989. At the time of publishing, this report is available on the web at: http://www.na.fs.fed.us/spfo/pubs/uf/sotuf/chapter_3/appendix_b/appendixb.htm Albers, J.; Hayes, E., How to Detect, Assess and Correct Hazard Trees in Recreational Areas, revised edition. St. Paul, MN: Minnesota DNR. 1993. 6-63

Construction and Maintenance Programs to Reduce Tree Faults

The USDA Forrest Service Northeastern Areas website at: http://www.na.fs.fed.us/

Right-of-Way Widening On many subtransmission lines, critical distribution lines, or circuit backbones, clearing a rightof-way is the most effective way to reduce the chance of tree contacts. Such right-of-ways are regularly maintained for high-voltage transmission lines. With normal distribution/subtransmission tree maintenance programs, many tree faults still occur. Even with hazard-tree programs, many tree faults will occur, either from healthy trees brought down by severe weather or from trees that die or are missed between maintenance cycles. The only way to drastically reduce tree faults is to clear a right-of-way. Then, the probability of a tree fault is determined by the width of the right-of-way and other factors, 199 including tree density, tree mortality rates, and tree heights. Guggenmoos outlines a methodology for estimating the risk of trees striking lines based on these factors. This approach can be used to estimate the benefit of a tree clearance program to establish a right-of-way or to widen an existing right-of-way. Audits Many utilities do audits after tree maintenance. Especially with contract crews, audits ensure that the work is being done to specifications. Even more so with hazard-tree programs and other targeted programs, audits can help educate tree crews at the same time that they ensure that the work is being done. Education comes from pointing out tree defects that were missed or tree cuts that should be made to reduce tree hazards or meet specified clearances. Utility Results With Targeted Programs Eastern Utilities Eastern Utilities, a small utility in Massachusetts (now a part of National Grid) implemented a danger-tree mitigation project with the following characteristics200201:

Three-phase primary circuits were targeted. Dead or structurally unsound trees were removed. Overhanging limbs were cut back. Trees were storm-proof pruned, meaning that trees were pruned to remove less severe structural defects. This was mostly crown thinning or reducing the height of a tree to reduce the sail effect.

Guggenmoos, S., "Effects of Tree Mortality on Power Line Security," Journal of Arboriculture, vol. 29, no. 4, pp. 181-96, July 2003a. 200 Simpson, P., "EUA's Dual Approach Reduces Tree-Caused Outages," Transmission & Distribution World, pp. 228, August 1997. 201 Simpson, P. and Van Bossuyt, R., "Tree-Caused Electric Outages," Journal of Arboriculture, May 1996.

199

6-64

Construction and Maintenance Programs to Reduce Tree Faults

On circuits where this was implemented, customer outage hours (SAIDI) due to tree faults were reduced by 20 to 30%. In addition, the program reduced tree-caused SAIDI by 62% per storm. Eastern Utilities did not increase funding for their vegetation management program to fund their danger-tree mitigation project. Instead, they funded the program by changes to their normal vegetation management program. They did less trimming of growth beneath the lines. They also embarked on a community communications effort to educate utility customers and win support for tree removal and more aggressive pruning. Also, they did not remove viable trees without the landowners consent. In addition, they found significant overall savings from reduced hot spotting and an even more significant savings from reduced outage restoration costs. Prior to implementing their program, Eastern Utilities surveyed random line sections to determine how extensive their program would need to be. They found that of the trees along those spans, 7% had excessive overhang, and another 6% were weak species or had a visible structural weakness. Niagara Mohawk After considerable study of tree-caused outages on their system, Niagara Mohawk Power Corporation implemented a program called TORO (Tree Outage Reduction Operation), which 202,203,204,205 had the following characteristics

Targeted work to the worst-performing circuits based on specific tree-caused indicators. Removed hazard trees located on targeted circuit segments. Specified greater clearances and removed overhanging limbs where possible on the backbone. Lengthened tree-maintenance cycles on rural 5-kV systems from 6 years to 7 or 8 years. Urban and suburban systems kept to a 5-year cycle. Looked for opportunities to improve system protection. They recently added inspection for the presence of single-phase tap fuses.

As of 2002, based on 250 feeders completed, on 92% of the feeders, tree SAIFI improved an average of 67%. More recent results show even more improvement.

EPRI 1008480, Electric Distribution Hazard Tree Risk Reduction Strategies, Electric Power Research Institute, Palo Alto, CA, 2004. 203 Finch, K., "Understanding Tree Outages," EEI Vegetation Managers Meeting, Palm Springs, CA, May 1 2001. 204 Finch, K., "Understanding Line Clearance and Tree Caused Outages," EEI Natural Resources Workshop, April 1, 2003a. 205 Finch, K. E., "Tree Caused Outages Understanding the Electrical Fault Pathway," EEI Fall 2003 Transmission, Distribution & Metering Conference, Jersey City, NJ, 2003b.

202

6-65

Construction and Maintenance Programs to Reduce Tree Faults

Puget Sound Energy Puget Sound Energy (PSE) implemented a hazard tree program they called TreeWatch206. Started in 1998, the focus of TreeWatch is on removing dead, dying, and diseased trees from private property along PSE's distribution system. On the circuits where they implemented their program, the average number of tree-caused outages and average outage duration dropped as shown in Table 6-13. They also found that they did not need to classify as many storms as major storms. Even in years with higher-than normal average windspeeds, PSE declared fewer storms as major storm events (where 5% of PSEs electric customers are without power due to weather-related causes).
Table 6-13 Results of the PSE TreeWatch Program Tree-Caused Outages TreeWatch Year 2001 (82 Circuits) 2000 (62 Circuits) 1999 (26 Circuits) Pre-TW 272 208 241 Post-TW 170 209 172 Change -37.4% 0.4% -28.6% Average SAIDI, min/yr Pre-TW 53.4 100.0 121.0 Post-TW 47.8 86.7 102.6 Change -10.4% -13.3% -15.2%

Source: Puget Sound Energy, "TreeWatch Program Annual Report," 2003. http://www.wutc.wa.gov/rms2.nsf/0/e25e1114a94abbf788256dd50082d992?OpenDocument

PSE started the TreeWatch program as a one-time program but hopes to continue the program at a reduced level of funding. Puget Sound Energy also estimates that they reduced the cost of treemaintenance on a per circuit mile basis by about 15%. They also estimated major reductions in storm restoration costs. Other Utility Programs Finch provides details on programs that ECI helped implement, including those by Niagara Mohawk (discussed previously), Kansas City Power & Light, and Flint EMC. KCPL and Flint EMC adjusted maintenance cycles to reduce cost and focus work on the most critical portions. On urban circuits, KCPL used a four-year cycle on the backbone with a two-year inspection to catch cycle busters and used a five-year cycle on laterals. On rural circuits, KCPL used a fiveyear cycle for all circuits. KCPL also developed a hazard-tree removal program based on results from their outage database. Flint EMC extended the maintenance cycle from four years to between five and six years on rural single-phase circuits.
207

Puget Sound Energy, "TreeWatch Program Annual Report," 2003. http://www.wutc.wa.gov/rms2.nsf/0/e25e1114a94abbf788256dd50082d992?OpenDocument 207 Finch, K. E., "Tree Caused Outages Understanding the Electrical Fault Pathway," EEI Fall 2003 Transmission, Distribution & Metering Conference, Jersey City, NJ, 2003b.

206

6-66

Construction and Maintenance Programs to Reduce Tree Faults

Another good resource is EPRI 1008480208, Electric Distribution Hazard Tree Risk Reduction Strategies. This documents hazard-tree program results by Niagara Mohawk, Central Hudson, and BC Hydro. They also provide a process tree map to help guide utilities through the process of developing a hazard-tree program.

EPRI 1008480, Electric Distribution Hazard Tree Risk Reduction Strategies, Electric Power Research Institute, Palo Alto, CA, 2004.

208

6-67

7
LIGHTNING PROTECTION, GROUNDING, AND ARRESTERS

Lightning Protection Background

Protection of power systems from lightning-related damage and faults is crucial to maintaining adequate power quality. In the United States, the most severe lightning activity occurs in the southeast and gulf coast states. However, lightning is also a major cause of faults and interruptions in areas with just modest lightning activity, such as New England. There are many facets in the design of lightning protection, including surge arrester sizing and placement, grounding issues, and the selection of appropriate power system equipment and insulation ratings. In the majority of cases, lightning causes temporary faults on distribution circuits; the lightning arcs externally across insulation, but neither the lightning nor the fault arc permanently damages any equipment. Normally, less than 20% of lightning strikes cause permanent damage. In Florida (high lightning), EPRI monitoring found permanent damage in 11% of circuit breaker operations that were coincident with lightning (EPRI TR-100218209 and Parrish210). Where equipment protection is done poorly, more equipment failures happen. Any failure in transformers, reclosers, cables, or other enclosed equipment causes permanent damage. One lightning flash may cause multiple flashovers and equipment failures. Even if a lightning-caused fault does no damage, a long-duration interruption occurs if the fault blows a fuse. The protection strategy at most utilities is:

Use surge arresters to protect transformers, cables, and other equipment susceptible to permanent damage from lightning. Use reclosing circuit breakers or reclosers to reenergize the circuit after a lightning-caused fault.

Lightning causes most damage by directly striking an overhead phase wire and injecting an enormous current surge that creates a very large voltage. The voltage impulse easily breaks down most distribution-class insulation unless it is protected with a surge arrester. Almost all direct lightning strokes cause flashovers. In addition, the lightning current may start a pole fire or burn

Characteristics of Lightning Surges on Distribution Lines. Second Phase, EPRI, Palo Alto, CA: 1991. TR100218. 210 Parrish, D. E., Lightning-Caused Distribution Circuit Breaker Operations, IEEE Transactions on Power Delivery, vol. 6, no. 4, pp. 1395401, October 1991.

209

7-1

Lightning Protection, Grounding, and Arresters

through conductors. Also, nearby lightning strokes that dont hit the line may couple damaging voltages to the line. These induced voltages may fail equipment or cause flashovers. Locations with higher lightning activity have higher fault rates on overhead circuits. Figure 7-1 shows fault rates reported by different utilities against estimated ground flash density. Utilities in higher lightning areas have more faults. Not all of these are due to lightning; many are due to wind and other storm-related faults. Lightning is a good indicator of storm activity at a location.

Fault rate per 100 miles per year

300

200

100

0 0 2 4 6 8 10

Ground flash density, flashes/km2/year

Figure 7-1 Distribution Line Fault Rate Versus Ground Flash Density

The linear curve fit to the data in Figure 7-1 shows line fault rates per 100 miles per year varying with the ground flash density, Ng, in flashes/km2/year as:
f = 20.6 N g + 43
Eq. 7-1

Direct Strikes When a cloud-to-ground lightning stroke hits an energized distribution conductor, a large amount of charge flows into the wire in a very, very short time period (microseconds). The flow of charge creates a voltage that is well above what normal distribution lines are designed to withstand. When this happens, the insulation between the energized wire and ground breaks down. Once the lightning has broken down the insulation, the power line electrical circuit has a short circuit. Short circuits caused by lightning occur at the point that is electrically the weakest, usually over the surface of an insulator at a pole location. Virtually all direct strokes cause faults. Lightning flashes hit typical-height distribution lines that are in the open at the rate of about 17 flashes/100 circuit miles/year for an area with Ng=1 flash/km2/year (11 flashes/100 km/year). Flashes to lines vary directly as the ground flash density. Almost all of these flashes to lines 7-2

Lightning Protection, Grounding, and Arresters

cause faults, and some damage equipment. Taller lines attract more flashes. The most accepted model of lightning strikes to power lines in open ground is Erikssons model, which finds the number of strikes as a function of the line height211:

28h 0.6 + b N = Ng 10
where: N is the flashes/100 km/year to the line Ng is the ground flash density per km2 per year ht is the height of the conductor (or overhead ground wire) at the tower in meters b is the overhead ground wire separation in meters For distribution lines, the b term can be ignored, which gives:
N = 2.8 N g h 0.6

Eq. 7-2

Eq. 7-3

for h in meters and N in flashes/100 km/year, or

N = 2.2 N g h 0.6
Eq. 7-4

for h in feet and N in flashes/100 miles/year For a 30-foot (10-m) line, about 17 flashes/100 miles/year (10.6 flashes/100 km/year) hit a 2 distribution line for Ng=1 flash/km /year. The main data point for distribution lines is a South African test line heavily monitored in the 1980s211. The test line had 19 flashes/100 miles/year (12 flashes/100 km/year) normalized for Ng=1 flash/km2/year for a line height of 28 feet (8.6 m). Erikssons equation is for a line in open groundmany lines have much fewer hits because of shielding from nearby objects (mainly trees, but also buildings and other power lines). For lines in forested areas, the number of strikes is significantly reduced, so this equation should be an upper limit for most lines. Unless some form of line protection is used (arresters or a shield wire), each of the direct strikes will cause a flashover and a fault on the system.

Eriksson, A. J., The Incidence of Lightning Strikes to Power Lines, IEEE Transactions on Power Delivery, vol. PWRD-2, no. 3, pp. 85970, July 1987.

211

7-3

Lightning Protection, Grounding, and Arresters

Induced Voltages Lightning strikes near a distribution line will induce voltages into the line from the electric and magnetic fields produced by the lightning stroke. These induced voltages are much less severe than direct strikes, but close strikes can induce enough voltage to flash insulation and damage poorly protected equipment. The charge and current flow through the lightning channel creates fields near the line. These fields induce voltages on the line. The vertical electric field is the major component that couples voltages into the line (magnetic fields also play a role). As the highly charged leader approaches the ground, the electric field increases greatly; and when the leader connects, the electric field collapses very quickly. The rapidly changing vertical electric field induces a voltage on a conductor, which is proportional to the height of the conductor above ground. The induced voltage waveform is usually a narrow pulse, less than five or ten microseconds wide, and it may be bipolar (negative then positive polarity). Most measurements of induced voltages have been less than 300 kV, so the most common guideline for eliminating problems with induced voltages is to make sure that the line insulation capability (CFO) is higher than 300 kV. Lines with insulation capabilities less than 150 kV have many more flashovers due to induced voltages. A simplified version of a model developed by Rusck212,213 approximates the peak voltage induced by nearby lightning. Rusck's model can be simplified to estimate the peak voltage developed on a conductor (IEEE Std. 1410-1997214):
V = 36.5 I h y

Eq. 7-5

where, V = peak induced voltage, kV I = peak stroke current, kA h = height of the line, usually feet or meters y= distance of the stroke from the line. This equation is for an ungrounded circuit; for a circuit with a grounded neutral or shield wire, the voltage from the phase to the neutral is less. For normal distribution line conductor spacings, multiply the answer by 0.75 for lines with a grounded neutral. So, for a 30-foot (10-m)
Rusck, S., Induced Lightning Overvoltages on Power Transmission Lines With Special Reference to the Overvoltage Protection of Low Voltage Networks, Transactions of the Royal Institute of Technology, no. 120, 1958. 213 Rusck, S., Protection of Distribution Lines in Lightning, R. H. Golde, Ed. London: Academic Press, 1977. 214 IEEE Std. 1410-1997, IEEE Guide for Improving the Lightning Performance of Electric Power Overhead Distribution Lines.
212

7-4

Lightning Protection, Grounding, and Arresters

distribution line, a 40-kA stroke 200 feet (30 m) from the line induces 165 kV on a line with a grounded neutral. Most lines are immune from strikes farther than 500 feet (150 m) from the line. Models of attraction to distribution lines show that for lines out in the open, most flashes that do not hit the line are too far away to induce a particularly high voltage across the insulation. For environmentally shielded lines, those with nearby trees and buildings, fewer strikes hit the line, but the line should have higher induced voltages because lightning strokes could hit closer to the line. EPRI sponsored rocket-triggered lightning tests in 1993 that showed induced voltages could be higher than predicted by Ruscks model for some strikes (Barker et al.215). For strikes to the ground 475 feet (145 m) from the line, voltages were 63% higher than Ruscks model. Hydro Quebec and New York State Electric and Gas sponsored another round of tests in 1994 that were also led by P. Barker. Closer strokes, strokes 60 feet (18 m) from the line, induced less voltage than the Rusck model. Table 7-1 compares the rocket-triggered lightning measurements with the Rusck model. High ground resistivity at the Florida test site probably explains why the 475-feet (145-m) measurements were higher than the Rusck prediction (Ishii216,217). Why the closer measurements are lower is not verified. One possible explanation is because those strokes were triggered from rockets launched on a 45-foot (14-m) tower. Because the tower is above ground, the positive charge on the tower shields some of the negative charge in the downward leader, so the electric field inducing voltages into the line is smaller. Environmentally shielded lines should act similarly to the tall towerthe charge collected on the tree should shield the line and reduce the induced voltage.

Barker, P. P., Short, T. A., Eybert, B. A. R., and Berlandis, J. P., Induced Voltage Measurements on an Experimental Distribution Line During Nearby Rocket Triggered Lightning Flashes, IEEE Transactions on Power Delivery, vol. 11, no. 2, pp. 98095, April 1996. 216 Ishii, M. Discussion to Barker et al., 1996 [2]. 217 Ishii, M., Michishita, K., Hongo, Y., and Oguma, S., Lightning-Induced Voltage on an Overhead Wire Dependent on Ground Conductivity, IEEE Transactions on Power Delivery, vol. 9, no. 1, pp. 10918, January 1994.

215

7-5

Lightning Protection, Grounding, and Arresters

Table 7-1 Comparison of Induced Voltage Measurements to Rusck Predictions Distance Number of Data Samples Rusck Prediction Vind/Is RocketTriggered Lightning Measurements Vind/Is 60 feet (18 m) 400 feet (125 m) 475 feet (145 m) 20 8 63 12.2 1.7 1.4 5.25 2.67 2.24

The three induced-voltage data points when normalized for a 30-foot (10-m) tall distribution line fit the following equation:
V= 9.8 I

1 + 121
y

1.8

Eq. 7-6

Figure 7-2 shows estimates of induced voltages with insulation level for both open ground and for lines that are environmentally shielded (usually by trees). One is based on Ruscks model using the approach from IEEE Std. 1410-1997. Another is based on the curve fit of the triggered lightning results, which shows more reasonable answers for environmentally shielded lines.

7-6

Lightning Protection, Grounding, and Arresters

100.0

Rusck model Rocket-triggered model

10.0

Flashovers/100 miles/year for Ng=1 flash/km2/year

Shielded lines

1.0

0.1

Lines in the open

0.01 0 100 200 300 400 500

CFO, kV

Figure 7-2 Induced Voltage Flashovers Versus Insulation Level for a Line With a Grounded Neutral

The main strategies for preventing induced flashovers are:

InsulationMaintain a 300-kV CFO on all structures. This eliminates most induced-voltage flashovers. Circuits with structures that have less than 100 kV of CFO are prone to many induced-voltage flashovers. ArrestersArresters at relatively wide spacings (every 6 structures) provide significant induced-voltage protection. Targeted arrester application at poles with poor insulation is even more effective at protecting against induced-voltage flashovers.

Surge Arresters and Arrester Reliability

Surge arresters are the primary defense against lightning damage on distribution circuits. Arresters must be placed as close as possible to the equipment being protected. This means on 7-7

Lightning Protection, Grounding, and Arresters

the same pole; because lightning has such high rates of rise, an arrester one pole span away provides little protection. At the pole structure, the best place for the arrester is right on the equipment tank to minimize the inductance of the arrester leads. Arresters protect any equipment susceptible to permanent damage from lightning. Transformers should have arresters. Two-terminal devices such as reclosers, regulators, and vacuum switches should have arresters on both the incoming and outgoing sides. Arresters on reclosers are especially importantwhen the recloser is open, a surge hitting the open point will double. Commonly, reclosers open to clear lightning-caused faults; and if the downstream side is not protected, a subsequent lightning stroke that followed the one that caused the fault could damage the open recloser. Cables need special attention to prevent lightning entry and damage. At equipment with predominantly air flashover pathsinsulators, switches, and cutoutsutilities do not normally use arresters. Make sure that arresters do not cause more problems than they solve. Arresters can cause problems, particularly:

Older arrestersOlder arresters on distribution systems are especially problematic. These older technologies include silicon carbide arresters, expulsion arresters, and other technologies. Animal faultsThe short length of arresters make them an easy target for squirrels and other animals.

Modern arresters are reliable components with failure rates much less than 1% annually. They do have somewhat of a bad reputation though, probably because there were bad products produced by some manufacturers during the early years of metal-oxide arresters. And, their tendency to fail violently did not help. One manufacturer states a failure rate less than 0.05% for polymer-housed arresters. This is primarily based on returned arresters, which are a low estimate of the true failure rate; arresters are almost a throw-away commodity; not all failures are returned. Arresters may also fail without utilities noticingif the isolator operates and a circuit breaker or recloser closes back in, the failed arrester is left with a dangling lead that may not be found for some time. Ontario Hydro 218 (CEA ) has recorded failure rates averaging about 0.15% in a moderately low-lightning area with about one flash/km2/year. About 40% of its arrester failures occurred during storm periods. Lightning arresters fail for a variety of reasons. Moisture ingress, failure due to lightning, and temporary overvoltages beyond arrester capability are some of the possibilities. Short et al.219 performed lab tests of arresters and EMTP modeling to evaluate system conditions such as ferroresonance, presence of distributed generation, and regulation voltages. Their main conclusion was that well-made arresters should perform well. More problems were likely for tightly applied arresters (such as using a 9-kV duty-cycle arrester with a 7.65-kV MCOV on a 12.47/7.2-kV system), so avoid tightly applying arresters under normal circumstances.
CEA 160 D 597, Effect of Lightning on the Operating Reliability of Distribution Systems, Canadian Electrical Association, Montreal, Quebec, 1998. 219 Short, T. A., Burke, J. J., and Mancao, R. T., Application of MOVs in the Distribution Environment, IEEE Transactions on Power Delivery, vol. 9, no. 1, pp. 293305, January 1994.
218

7-8

Lightning Protection, Grounding, and Arresters

Darveniza et al.220 inspected several arresters damaged in service in Australia. Of the gapped metal-oxide arresters inspected, both polymer and porcelain-housed units, moisture ingress caused most failures. Of the gapless metal-oxide arresters inspected, several were damaged under conditions that were likely ferroresonancearresters on riser poles with single-pole switching or with a cable fault. Another portion was likely damaged from severe lightning, most likely from multiple strokes. Most of the failures occurred along the outer surface of the blocks. A small number of metal-oxide arresters, both polymer and porcelain, showed signs of moisture ingress damage, but overall, the metal-oxide arresters had fewer problems with moisture ingress than might be expected. Lightning causes some arrester failures. The standard test waves (4/10 s or 8/20 s) do not replicate lightning very well but are assumed to test an arrester well enough to verify field performance. The energy of standard test waves along with the charge is shown in Table 7-2. Charge corresponds well with arrester energy input because the arrester discharge voltage stays fairly constant with current.
Table 7-2 Approximate Energy and Charge in ANSI/IEEE Arrester Test Current Waves Test Wave 100 kA, 4/10 s 40 kA, 8/20 s 10 kA, 8/20 s Energy, kJ/kV of MCOV 4.5 3.6 0.9 Charge, Coulombs 1.0 0.8 0.2

Studies of surge currents through arresters show that individual arresters conduct only a portion of the lightning current. Normally, the lightning current takes more than one path to ground. Often that path is a flashover caused by the lightningthe flashover is a low-impedance path that protects the arrester. The largest stroke through an arrester measured during an EPRI study with more than 200 arrester years of monitoring using lightning transient recorders measured 28 221 kA (Barker et al. ). During the study, 2% of arresters discharged more than 20 kA annually. The largest energy event was 10.2 kJ per kV of MCOV rating (the arrester did not fail, but largerthan-normal 4.7-cm diameter blocks were used). Manufacturers often cite 2.2 kJ per kV of MCOV rating as the energy capability of heavy-duty arresters. The actual capability is probably higher than this. If test results from station-class arresters (Ringler et al.222) are translated to equivalent distribution size blocks, heavy-duty arresters should withstand 6 to 15 kJ/kV of MCOV (the 100-kA test produces about 4 to 7 kJ/kV
Darveniza, M., Saha, T. K., and Wright, S., Comparisons of In-Service and Laboratory Failure Modes of Metaloxide Distribution Surge Arresters, IEEE/PES Winter Power Meeting, October 2000. 221 Barker, P. P., Mancao, R. T., Kvaltine, D. J., and Parrish, D. E., Characteristics of Lightning Surges Measured at Metal Oxide Distribution Arresters, IEEE Transactions on Power Delivery, vol. 8, no. 1, pp. 30110, January 1993. 222 Ringler, K. G., Kirkby, P., Erven, C. C., Lat, M. V., and Malkiewicz, T. A., The Energy Absorption Capability and Time-to-Failure of Varistors Used in Station-Class Metal-Oxide Surge Arresters, IEEE Transactions on Power Delivery, vol. 12, no. 1, pp. 20312, January 1997.
220

7-9

Lightning Protection, Grounding, and Arresters

of MCOV). Metal-oxide blocks exhibit high variability in energy capability, even in the same manufacturing batch. Darveniza et. al.223found that multiple strokes are more damaging to arresters. Less energy is needed to fail the arrester. The failures occur as flashovers along the surface of the metal-oxide blocks. The block surface coating plays an important rolecontamination and moisture increased the probability of failure. Five 8/20-s test waves were applied within an interval of 40 ms between each. A set of five 10-kA impulses applies about 4 kJ/kV of MCOV. Many of the arrester designs failed these tests. The best design had a housing directly molded onto the blocks. Longer-duration surges also fail arresters at lower energy levels. Tests by Kannus et. al.224 on polymer-housed arresters found many failures with 2.5/70-s test waves with a peak value of about 2 kA. As with multiple impulse tests, the outside of the blocks flashed over. Some arresters failed with as little as 0.5 kJ/kV of MCOV. Moisture ingress increased the failure probability for some arrester designs (Lahti et al.225). Older Arresters Many older arrester technologies are still in place on distribution circuits (see Figure 7-3 for examples). Many older distribution circuits have a variety of arrester technologies, including silicon-carbide arresters, thyrite arresters, pellet arresters, and expulsion arresters. Older arrester technologies all have an internal gap. The gap can be a source of repeated power quality problems. The gap can sparkover because of moisture. After the gap sparks over, power follow current will flow through the arrester; the amount of current depends on the materialanything from hundreds of amps in a silicon-carbide arrester to the bolted short-circuit level at the location in an expulsion arrester (more nonlinear arrester materials have less power follow current). The power follow current can blow fuses or trip other protective devices. Even on an arrester that is completely failed to a short circuit, the gap may be able to hold full system voltage until the gap sparks over.

Darveniza, M., Roby, D., and Tumma, L. R., Laboratory and Analytical Studies of the Effects of Multipulse Lightning Current on Metal Oxide Arresters, IEEE Transactions on Power Delivery, vol. 9, no. 2, pp. 76471, April 1994. 224 Kannus, K., Lahti, K., and Nousiainen, K., Effects of Impulse Current Stresses on the Durability and Protection Performance of Metal Oxide Surge Arresters, High Voltage Engineering Symposium, Aug. 2227, 1999. 225 Lahti, K., Kannus, K., and Nousiainen, K., The Durability and Performance of Polymer Housed Metal Oxide Surge Arresters Under Impulse Current Stresses, Cired, 2001.

223

7-10

Lightning Protection, Grounding, and Arresters

Expulsion Arrester Figure 7-3 Examples of Older Arresters

Peroxide-Pellet Arrester

Silicon-Carbide Arrester

The most prevalent type of arrester is the gapped silicon-carbide arrester. Silicon-carbide is a nonlinear resistive material, but it is not as nonlinear as metal oxide. It requires a gap to isolate the arrester under normal operating voltage. When an impulse sparks the gap, the resistance of the silicon-carbide drops, conducting the impulse current to ground. With the gap sparked over, the arrester continues to conduct 100 to 300 A of power follow current until the gap clears. If the gap fails to clear, the arrester will fail. Annual failure rates have been about 1% with moisture ingress into the housing causing most failures of these arresters (an Ontario-Hydro survey found 226 that 86% of failures were from moisture (Lat and Kortschinski ). Moisture degrades the gap, damaging it outright or preventing it from clearing a surge properly. Darveniza et. al.227 recommended that gapped silicon-carbide arresters older than 15 years be progressively replaced with metal-oxide arresters. Their examinations and tests found that a significant portion of silicon-carbide arresters had serious deterioration with a pronounced upturn after about 13 years. Externally gapped arresters can especially be a source of repeated problems (Figure 7-4). If the arrester is failed, all it takes is an animal contact (or other debris) across a very short gap to cause a fault on the system. Even insects have been noted to cause faults in such situations.
226

Lat, M. V., and Kortschinski, J., Distribution Arrester Research, IEEE Transactions on Power Apparatus and Systems, vol. PAS-100, no. 7, pp. 3496505, July 1981. 227 Darveniza, M., Mercer, D. R., and Watson, R. M., An Assessment of the Reliability of In-Service Gapped Silicon-Carbide Distribution Surge Arresters, IEEE Transactions on Power Delivery, vol. 11, no. 4, pp. 178997, October 1996.

7-11

Lightning Protection, Grounding, and Arresters

Use every opportunity to replace older arresters. Doing so will remove threats to power quality from repeated flashovers and improve the protective level provided by modern arresters.

Source: EPRI 1001883

Figure 7-4 Example of a Gapped Arrester on a Recloser

Arrester Installation Considerations Distribution arresters have isolators that remove failed arresters from the circuit. The isolator has an explosive cartridge that blows the end off of a failed arrester, which provides an external indication of failure. The isolator itself is not designed to clear the faultan upstream protective device normally must clear the fault (although in a few cases, the isolator may clear the fault on its own, depending on the available short-circuit current and other parameters). Crews should take care with the end lead that attaches to the bottom of the arresterit should not be mounted such that the isolator could swing the lead into an energized conductor if the isolator operates. The proper lead size should also be used (make sure that it is not too stiff, which might prevent the lead from dropping). Figure 7-5 shows an example of an arrester lead that has dropped dangerously close to the energized bushing of a transformer. An animal contact could easily cause a line-to-ground fault in this situation.

7-12

Lightning Protection, Grounding, and Arresters

Tight spacing

Figure 7-5 Blown Arrester With a Dangling Ground Lead

Arrester versus fuse placement is an ongoing industry debate. Tank mounting an arrester protects the transformer best; but because the transformer is downstream of the fuse, the lightning surge current passes through the fuse. Lightning may blow the fuse unnecessarily, and many utilities have histories of nuisance fuse operations. Applying the arrester upstream of the fuse keeps the surge current out of the fuse, but usually results in long lead lengths. Overall, the tank-mounted approach is best along with using larger fuses or surge-resistant fuses to limit unnecessary fuse operations. Completely self-protected transformers have an internal fuse rather than an external fuse. Because of this, if the arrester fails, a tap fuse or upstream recloser or even the substation breaker will have to trip to clear the fault. Many more customers will be affected. If the arrester isolator fails to operate (which can happen), the failure may be extremely hard for the crews to find. Line-protection arrestersthose installed to provide extra protection to the line itselfare another source of problems. In addition to adding a possible animal contact to the system, if the arrester fails, a large number of customers may have an interruption. Such installations are not normally fused. Figure 7-6 shows an example of a blown line-protection arrester. If lineprotection arresters are used, consider fusing as a way of isolating arrester failures from affecting more customers.

7-13

Lightning Protection, Grounding, and Arresters

Figure 7-6 also points out that when crews apply arresters, they should consider the possibility that the phase end of the arrester may also come apart if the arrester fails. This is harder than preventing errant ground leads from causing further problems because phase jumper leads can be quite long.

Figure 7-6 Example of a Blown Arrester

Pole Insulation and Lightning

High insulation levels on structures help prevent induced voltage flashovers. Insulation levels are also critical on some types of line-protection configurations with shield wires or arresters. Note that for a normal distribution line, higher insulation levels may actually stress nearby cables and transformers more. With high insulation strengths, a higher voltage develops across the insulation before it flashes over. By allowing a larger magnitude surge on the line before flashover, damage to nearby equipment is more likely. The flashover of the insulation acts as an arrester and protects other equipment. The critical flashover voltage (CFO) of self-restoring insulation (meaning no damage after a flashover) is the voltage where the insulation has a 50% probability of flashing over for a standard 1.2/50-s voltage wave. For insulators, manufacturers catalogs specify the CFO. CFO and BIL (Basic Lightning Impulse Insulation Level) are often used interchangeably, but they have slightly different definitions. A statistical BIL is the 10% probability of flashing over for a standard test wave. Normally, CFO and statistical BIL are within a few percent of each other. Lightning may flash along several paths, directly between conductors across an air gap or along the surface of insulators and other hardware at poles (normally the easiest path to flashover). We need to consider phase-to-ground and phase-to-phase paths. At a pole structure, the flashover path may involve several insulating components, the insulator, wood pole or crossarm, and possibly fiberglass. Wood and fiberglass greatly increase the structure insulation. 7-14

Lightning Protection, Grounding, and Arresters

Table 7-3 shows the critical flashover voltage of common components. When more than one insulator is in series, the total insulating capability is less than the sum of the components. When insulation is subjected to a voltage surge, the voltage across each component splits based on the capacitance between each element. Normally, this voltage division does not split the voltage by the same ratio as their insulation capability, so one component flashes first leaving more voltage across the rest of the components.
Table 7-3 CFO of Common Distribution Components (by Themselves) Component Air Wood pole or crossarm Fiberglass standoff kV/foot 180 100 150 kV/m 600 350 500

The simplest way to estimate the insulation level of a structure is to take the CFO of the insulator (usually about 100 kV) and add the wood length at 75 kV per foot (250 kV/m). Often the wood provides more insulation than the insulator. Estimate the air-to-air gap using 180 kV/foot (600 kV/m). The air gap between conductors usually has higher insulation than the path along the insulator and wood. So, most flashovers occur at the poles, the weakest point. Typical distribution structures generally have CFOs between 150 and 300 kV. And, to eliminate induced voltage flashovers, we try to have 300 kV of CFO. On wood structures, this means having at least 2.5 feet (0.75 m) of wood along with all possible flashover paths. Another way to estimate the structure CFO of several components in series is to take the square root of the sum of the squares of each component. Another more precise way to estimate 228 structure CFOs is with the extended CFO added method described by Jacob et al. and Ross and Grzybowski229 (or, see IEEE Std. 1410-1997 for a simplified version). Insulation and Pole Construction Guidelines Weak-link polesequipment poles, crossover structures, or other busy structureswith low insulation significantly increase induced-voltage flashovers. The following components can severely degrade pole structure insulation:

228

Fuse cutouts Metal braces and supports Ground wires near phase conductors Grounded or poorly insulated guy insulators

Jacob, P. B., Grzybowski, S., and Ross, E. R. J., An Estimation of Lightning Insulation Level of Overhead Distribution Lines, IEEE Transactions on Power Delivery, vol. 6, no. 1, pp. 38490, January 1991. 229 Ross, E. R., and Grzybowski, S., Application of the Extended CFO Added Method to Overhead Distribution Configurations, IEEE Transactions on Power Delivery, vol. 6, no. 4, pp. 15738, October 1991.

7-15

Lightning Protection, Grounding, and Arresters

Fuse cutouts are often attached such that a lower-than-normal flashover path is available. Arrange the cutout so that the attachment bracket is mounted on the pole away from any grounds, including guy wires. It is difficult to maintain high insulation capabilities on poles with multiple circuits. Distribution circuits built under transmission lines have added lightning exposure. A transmission line is taller and attracts more strokes, and usually a down ground runs just by the distribution line (leaving little wood insulation between the distribution insulation and ground). Armless designs have less insulating capability (see Figure 7-7), especially if metal insulator spacers are used (fiberglass arms are better for compact spacings). Spacer cables also have low insulation (CFOs are usually 150 to 200 kV). Spacer cables have a messenger cable that acts as a shield wire, but the insulating capability is so low that it barely provides benefit unless grounding is outstanding.

Courtesy: Duke Power

Figure 7-7 Example of an Armless Design

Do not use pole-protection assemblies (see Figure 7-8). This design has a pole bonding wire that runs up to the base of the top insulator and has a four-inch (10-cm) spark gap between the bonding wire and the phase conductor. This assembly was recommended by the Rural Electric Association at one time for locations where conditions indicate that pole damage may be a problem. The CFO is about 125 kV, which is so low that it leads to many induced-voltage flashovers. Figure 7-9 shows additional examples where ground wires are run up near the primary. In most cases this is unnecessary. Another situation where ground wires are run near primary wires is when guy wires are grounded.

7-16

Lightning Protection, Grounding, and Arresters

Pole-protection assembly

Insulator damage

Courtesy: Duke Power

Figure 7-8 Example Pole-Protection Assemblies

7-17

Lightning Protection, Grounding, and Arresters

Courtesy: Duke Power

Figure 7-9 Example of Wood Poles Bonded With Ground Wires

Although service experience indicates that lightning rarely damages poles or crossarms, in highlightning areas, concern is warranted. Darveniza230 cites a survey by Zastrow published in 1966 that found poles failing from lightning in the range of 0.008 to 0.023% annually (versus an overall failure rate of 0.344 to 1.074% with more than half of these due to decay). Other surveys summarized by Darveniza also found minimal lightning-caused damage. Lightning overvoltages damage and shatter poles and crossarms when the wood breaks down internally rather than along the surface. If the wood is green and wet, internal breakdowns and damage are more likely. Damage within the first year of service is more likely. Figure 7-10 shows examples of poles probably damaged by lightning.

230

Darveniza, M., Electrical Properties of Wood and Line Design, University of Queensland Press, 1980.

7-18

Lightning Protection, Grounding, and Arresters

Figure 7-10 Examples of Poles With Probable Lightning Damage

If historical records show that wood damage is a problem, bonding the insulators (grounding the base of each insulator) protects the wood, but this shorts out the insulation capability provided by the wood. Better solutions are surface electrodes fitted near the insulator pin, including wirewraps, bands, or other metal extensions attached near the insulator in the likely direction of flashover. This local bonding encourages breakdown near the surface rather than internally. Local bonding is better than completely bonding the insulators because the insulation level is not compromised. Poorly insulated guy wires near phase conductors are a source of lightning problems (as well as animal problems). Grounded guy wires near phase conductors provide a weak flashover point (Figure 7-11). Use fiberglass guy-strain insulators rather than the small porcelain guy-strain insulators. Guys are often the lowest flashover path because they are attached right near the top insulator.

7-19

Lightning Protection, Grounding, and Arresters

Courtesy: Duke Power

Figure 7-11 Example of Uninsulated Guy Wires Near Phase Conductors

7-20

Lightning Protection, Grounding, and Arresters

Figure 7-12 Example of a Poorly Insulated Guy Wire Near Phase Conductors

Fiberglass has excellent insulating properties, but crews must not short out the insulation. See Figure 7-13 for a humorous example where crews used a very long fiberglass guy-strain insulator and then wrapped a ground lead around it, thereby shorting out all of that insulation capability.

Fiberglass insulator shorted out

Courtesy: Duke Power

Figure 7-13 A Fiberglass Guy Insulator Shorted Out

7-21

Lightning Protection, Grounding, and Arresters

Arresters are another approach to dealing with insulation at weak-links poles. Rather than enhancing the insulation at weak-link poles, add arresters to the pole. Underbuilt lines and weaklink pole structures can be protected against induced-voltage flashovers with arresters.

Line Protection
Line protection is the attempt to reduce the number of lightning-caused faults. Utilities have increased interest in line protection as one way to improve reliability and power quality. Because lightning can flash over a 230-kV or 500-kV transmission line, we should not be surprised that protecting a 13-kV distribution line is difficult. To protect against direct hits, we need either a shield wire or arresters to divert the stroke to ground without a flashover. Lightning strokes close to a line may induce enough voltage to flash over the line insulation. Induced voltages are easier to contain because induced voltages have much lower magnitudes than direct-strike voltages. Maintaining enough insulation capability is the normal way to limit induced-voltage flashovers. Line arresters can also greatly reduce induced-voltage flashovers from nearby strokes. Shield Wires Shield wires are effective for transmission lines but are difficult to make work for distribution lines. A shield wire system works by intercepting all lightning strokes and providing a path to ground. If the path to ground is not good enough, a voltage develops on the ground with respect to the phases (called a ground potential rise). If this is high enough, the phase can flashover (called a backflashover). Grounding and insulation are important. Good grounding reduces the ground potential rise. Extra insulation protects against backflashover. As an example, consider Figure 7-14, where a 22-kA stroke (which is on the small size for lightning) is hitting a distribution line. The ground potential rises to 400 kV relative to the phase conductor, enough voltage to flashover most distribution lines.

7-22

Lightning Protection, Grounding, and Arresters

22 kA

Shield wire 1 kA

1 kA

Phase wire V = 20 kA(20 ohms) = 400 kV

20 kA

20 ohms

Figure 7-14 Shield-Wire Lightning Protection System

100

Percentage of direct strikes that cause flashovers

80

100 kV
60 40 20 0 10.0

200 kV

300 kV CFO

20.

50.

100.0

200.

500.

Pole footing resistance, ohms

Figure 7-15 Performance of a Shield Wire Depending on Grounding and Insulation Level

7-23

Lightning Protection, Grounding, and Arresters

Shield wire Phase wire Shielding angle

Figure 7-16 Shield Wire Shielding Angle

To keep the insulation high, use fiberglass standoffs to keep the ground wire away from the pole to maximize the wood length. Also, make sure guy wires and other hardware do not compromise insulation. Ground the shield wire at each pole. Lightning has such fast rise times that if a pole is not grounded and a lightning strike hits the shield wire, voltage will build up so fast at the ungrounded pole that the insulation will flash over before reflections from adjacent grounded poles can provide any help in relieving the voltage stress. At all poles, obtain a ground that is 20 or less. Good grounding is vital. Exposed sections of circuit such as at the top of a hill or ridge should have the most attention. Getting adequate grounds may require:

More than one ground rodmake sure to keep them spaced further than one ground rod apart. Deeper ground rods. Chemical soil treatments. Counterpoise wires (buried lengths of wire).

Figure 7-15 shows estimates of performance versus grounding for several insulation levels based on the approach of IEEE Std. 1410-1997. In order to ensure that lightning hits the shield wire and not the phase conductors, maintain a shielding angle of 45 or less (as defined in Figure 7-16). Line Protection Arresters Arresters are normally used to protect equipment. Some utilities are using them to protect lines against faults, interruptions, and voltage sags. To do this, arresters are mounted on poles and attached to each phase. For protection against direct strikes, arresters must be spaced at every pole (or possibly every other pole on structures with high insulation levels) (McDermott et al.231).
231

McDermott, T. E., Short, T. A., and Anderson, J. G., Lightning Protection of Distribution Lines, IEEE Transactions on Power Delivery, vol. 9, no. 1, pp. 138152, January 1994.

7-24

Lightning Protection, Grounding, and Arresters

This is a lot of arresters, and the cost prohibits widespread usage. The cost can only be justified for certain sections of line that affect important customers. Arresters have been used at wider spacings such as every four to six poles by utilities in the southeast for several years. This grew out of some work done in the 1960s by a task force of 232,233 eight utilities and the General Electric Company . Anecdotal reports suggest improvement, but there is little hard evidence. Recent field monitoring and modeling suggest that this should not be effective for direct strikes. One of the reasons that this may provide some improvement is that arresters at wider spacings improve protection against induced voltages. Nevertheless, arresters applied at a given spacing are not recommended as the first optionfixing insulation problems or selectively applying arresters at poles with poor insulation are better options for reducing induced-voltage flashovers. For direct-strike protection, arresters are needed on all poles and on all phases. The amount of protection quickly drops if wider spacings are used. Lead length is not as much of an issue as it is with equipment protection, but it is always good practice to keep lead lengths short. The arrester rating would normally be the same as the existing arresters. Grounding is normally not an overriding concern if arresters are used on all phases. If grounds are poor, one effect is that surges tend to get pushed out away from the strike location (because there is no good path to ground). One of the implications of this is if just a section of line is protected (such as an exposed ridge crossing), then grounding at the ends is important. Good grounds at the end provide a path to drain off the surge. One concern with arresters is that they may have a relatively high failure rate. Direct strikes can cause failures of nearby arresters. Something in the range of 5 to 30% of direct lightning strikes may cause an arrester failure. This is still an undecided (and controversial) subject within the industry. It is recommended that the largest block size available be used (heavy-duty or intermediate-class blocks) to reduce the probability of failure. Field trials on Long Island, NY, of arresters at various spacings did not show particularly 234 promising results for distribution line protection arresters (Short and Ammon ). LILCO (Long Island Lighting Company) added line arresters to three circuits. One had spacings of 10 to 12 spans between arresters (1300 feet, 400 m), one had spacings of five to six spans (600 feet, 200 m), and one had arresters at every pole (130 feet, 40 m). Arresters were added on all three phases. We also monitored two other circuits for comparison. None of the three circuits with line arresters had dramatically better lightning-caused fault rates than the two circuits without arresters. Statistically, we cannot infer much more than this because the data are limited (always a problem with lightning studies). One of the most significant results was that the circuit with arresters on every pole had several lightning-caused interruptions, and theoretically it should
Task force of eight utility companies and the General Electric Company, Investigation and Evaluation of Lightning Protective Methods for Distribution Circuit. I. Model Study and Analysis, IEEE Transactions on Power Apparatus and Systems, vol. PAS-88, no. 8, pp. 12328, August 1969. 233 Task force of eight utility companies and the General Electric Company, Investigation and Evaluation of Lightning Protective Methods for Distribution Circuits. II. Application and Evaluation, IEEE Transactions on Power Apparatus and Systems, vol. PAS-88, no. 8, pp. 123947, August 1969. 234 Short, T. A., and Ammon, R. H., Monitoring Results of the Effectiveness of Surge Arrester Spacings on Distribution Line Protection, IEEE Transactions on Power Delivery, vol. 14, no. 3, pp. 11421150, July 1999.
232

7-25

Lightning Protection, Grounding, and Arresters

have had none. Missing arresters is the most likely reason for most of the lightning-caused faults. One positive result was that the arrester failure rate was low on the circuit with arresters on every pole. Another field study showed more promise for line arresters. Commonwealth Edison added 235 arresters to several rural, open feeders in Illinois (McDaniel ). ComEd uses an arrester spacing of 1200 feet (360 m); as a trial, they tightened the arrester spacing to 600 feet (180 m) on 30 feeders (and all existing arresters that were not metal oxide were replaced). The 30 feeders with the new spacing were compared to 30 other feeders that were left with the old standard. Over three lightning seasons of evaluation, the upgraded circuits showed that circuit interruptions improved 16% (at a 95% confidence level). Note that most of the interruptions were transformer fuse operations. Another way to apply arresters is to use them on the top phase only. The top phase is turned into a shield wire. When lightning hits the top phase, the arrester conducts and provides a lowimpedance path to the pole ground. Just as with a shield-wire system, grounding and insulation are critical. A top-phase arrester application cannot be used on typical three-phase crossarm designs because there is no top phase. In areas where grounding is poor and arrester failure is a concern, arresters can be used with a shield wire system. The shield wire takes away most of the energy concerns, and the arresters protect against backflashovers. This provides very good lightning protection (but is very expensive).

Grounding
Many utility workers have misconceptions about grounding. Poor grounding is blamed for many problems, whether true or not. Many conceptions about grounding are based on anecdotal evidence. Related to lightning-caused faults, grounding can affect performance in a number of ways:

Equipment damage rateImproved grounding will help reduce equipment damage on distribution circuits. Better grounding confines a surge to the vicinity of the strike-point; if grounding is poor, a voltage surge can propagate further out and subject more equipment to possibly damaging surges. Better grounding will reduce distribution transformer failures by reducing failures from secondary surge entry. Line fault rateImproved grounding may not change the overall fault rate on a circuit. Direct strikes will cause flashovers regardless of grounding (unless shield wires or arresters are used for line protection). Grounding effects on induced voltages are complicated: higher ground resistivities lead to higher magnitude induced voltages, but better pole grounding can actually increase phase-to-neutral voltages. Overall, circuits with poor grounding may see more damage, but not necessarily more faults. Line interruption rateWhile a circuit with better grounding may not have a better fault rate, it may have a better interruption rate, mainly because fewer fuses should blow. Poor

235

McDaniel, J., Line Arrester Application Field Study, IEEE/PES Transmission and Distribution Conference and Exposition, 2001.

7-26

Lightning Protection, Grounding, and Arresters

grounding may increase the equipment damage rate (more long-duration interruptions). Poor grounding can also cause cascading faults that can blow multiple tap fuses.

Line protection arrestersBetter grounding can help in certain applications where arresters are used to provide line protection. Shield wire applicationOn configurations with a shield wire, good grounding is critical for protecting the circuit against backflashovers from direct strikes. Customer surgesPoor grounding can increase lightning surges that enter into customer facilities (both residential and commercial). Poor grounding may also force more surge current into telephone and cable television systems.

With poor grounding, lightning strikes to distribution line subject more equipment to possibly damaging surges due to ground potential rise. Lightning current must flow to ground somewhere; if the pole ground near the strike point has high impedance, more of the surge flows in the line, exposing more equipment to possible overvoltages. At the point of a direct stroke to a distribution conductor, the huge voltages flash over the insulation, shorting the phase to the neutral. From there, the lightning travels to the closest ground; if there is no proper pole ground nearby, the next most likely path is down a guy wire. If the path to ground is poor, the phase conductor and neutral wire all rise up in voltage together. This surge (on both the phase and neutral) travels down the circuit. When the surge reaches another ground point, current drains off the neutral wire, which increases the voltage between the phase and the neutral. The voltage across the insulation grows until a surge arrester conducts more current to ground or another insulation flashover occurs. Figure 7-17 shows a drawing describing the ground potential effect.
30 kA V = 5 kA(400 ohms) = 2000 kV 5 kA Phase wire flashover V = 0 LN Neutral wire 5 kA 10 kA 5 kA V = 10 kA(200 ohms) = 2000 kV 200 ohms 5 kA V = 5 kA(10 ohms) = 50 kV 10 ohms VLN = 2000 kV - 50 kV = 1950 kV

5 kA

Figure 7-17 Impact of Grounding

Good grounding helps confine possible lightning damage to the immediate vicinity of the strike. If the distribution insulation flashes over, the short circuit acts as an arrester and helps protect other equipment on the circuit (as long as grounds are good). Normally, the ground resistance right at a piece of equipment protected with an arrester does not impact the primary-side protection. The primary surge arrester is between the phase and ground. For a lightning hit to the primary, the arrester conducts the current to ground and limits the voltage across the equipment insulation (even though the potential of all conductors may rise with poor grounding). 7-27

Lightning Protection, Grounding, and Arresters

Grounding does play a role for transformers, which are vulnerable to surge current entering through the secondary. Poor grounding pushes more current into the transformer on the secondary side and increases the possibility of failure. Poor grounding also forces more current to flow into the secondary system to houses connected to the transformer. And, poor grounding may also push more current into telephone and cable television systems. Secondary-Side Surges and Grounding Single-phase residential-type transformers (three-wire 120/240-V service) may also fail from surge entry into the low-voltage winding. This damage mode has been extensively discussed within the industry (sparking competing papers from manufacturers) and summarized by an IEEE Task Force236. Lightning current into the neutral winding of the transformer (usually X2) induces possibly damaging stresses in the high-voltage winding near the ground and line ends, both turn-to-turn and layer-to-layer. Lightning current can enter X2 from strikes to the secondary or strikes to the primary (where it gets to the neutral through a surge arrester or flashover, as shown in Figure 7-18).

IEEE Task Force Report, Secondary (Low-Side) Surges in Distribution Transformers, IEEE Transactions on Power Delivery, vol. 7, no. 2, pp. 74656, April 1992.

236

7-28

Lightning Protection, Grounding, and Arresters

No Customer Load

Loop voltages induced

Pole ground

Service entrance ground

With Load (or meter gap sparkover)

Pole ground

Service entrance ground

Figure 7-18 Lightning Surge Entry via the Secondary

The most concern with secondary surge entry is on small (< 50 kVA) overhead transformers with a non-interlaced secondary winding (pad-mounted transformers are also vulnerable). The primary arrester does not limit the voltage stresses on the ends of the primary winding. Currents couple less to the high-voltage winding on transformers with an interlaced secondary winding, which is used on core-type transformers and some shell-type transformers. Smaller kVA transformers are most prone to damage from this surge entry mode. When the surge current gets to the transformer neutral, it has four places to go: down the pole ground, along the primary neutral, along the secondary neutral towards houses, or into the transformer winding. Current through the secondary neutral creates a voltage drop along the neutral. This voltage drop will push current through the transformer when load is connected at the customer (Figure 7-18). The voltage created by flow through the neutral is partially offset by the mutual inductance between the secondary neutral and the phases (tighter coupling is better). As far as the effects on the transformer, the major factors impacting secondary-surge entry into the transformer are:

7-29

Lightning Protection, Grounding, and Arresters

InterlacingFor a balanced surge (equal in both windings), the inductance provided by the transformer is approximately the hot leg to hot leg short-circuit reactance. Transformers with interlaced secondaries have significantly lower inductances. Transformer sizeSmaller transformers have higher inductances, inducing a higher voltage on the primary-side winding. GroundingBetter grounding on the transformer helps reduce the current into the transformer. Poor grounding sends more current into the transformer and more current into houses connected to the transformer. Other nearby primary grounds help to reduce the current into the transformer. Secondary wireSince triplex has tighter coupling than open-wire secondaries (and less voltage induced in the loop), a secondary of triplex has less current into the transformer neutral terminal. Multiple secondariesAlso, the worst case is with one secondary drop from the transformer; several in parallel reduce the effective impedance of the secondary. Secondary lengthLonger secondaries are more prone to transformer failure (higher inductance). Similar factors affect the magnitudes of surges into residential customers: GroundingBetter grounding on the transformer helps reduce the current into customer facilities. Multiple secondariesAgain, the worst case is with one secondary drop from the transformer; several in parallel reduce the effective impedance of the secondary. Transformer interlacing and transformer secondary arrestersUsing interlaced transformers helps protect the transformer but can increase the current sent into the residential customer. The same effect holds for secondary arrestersthey provide better protection for the transformer but force more current into the house.

Summary of Guidelines for Lightning Protection, Grounding, and Arrester Application

The main recommendations for applying arresters are:

Equipment applicationIn all but the very lowest lightning areas, use lightning arresters at all equipment poles, including transformers, riser poles, capacitors, reclosers, and regulators. Line applicationMany utilities in high-lightning areas install line-protection arresters at mile (0.4-km) spacings to improve line performance. These spacings will help protect against induced voltages and will help reduce equipment failures; they will not help reduce directstroke flashovers. The downside to line-protection arresters is that they add another component to the system that can fail and add another location where animal faults can occur. Animal protectionBecause of their short line-to-ground length, arresters are a prime location for animal-caused faults. Use animal guards on all arresters.

7-30

Lightning Protection, Grounding, and Arresters

Older arrestersReplace older arresters wherever possible. Externally gapped arresters are a significant source of power quality problems; they can fail internally and cause repeated voltage sags and fuse operations. FusingApply arresters downstream of local fuses wherever possible. Leaving an arrester unprotected means that the next upstream device (tap fuse, recloser, or circuit breaker) must clear the fault. More customers are interrupted, the duration can be longer (the problem is harder to find), and the voltage sag is more severe. MountingMount arresters such that arrester failures do not impact other phases. Use the appropriate ground lead, and mount the arrester and ground lead so that the ground lead cannot swing into other energized conductors if the arrester isolator operates.

The main goal on structures is to maintain a sufficient insulation capability to avoid most induced voltage flashovers. The goal is a 300-kV critical flashover voltage. On wood structures, have at least 2.5 feet (0.75 m) of wood along all possible flashover paths. Particularly, consider the following advice to eliminate weak-link structures that may be particularly prone to flashover:

Guy wiresGrounded guy wires near phase conductors provide a weak flashover point. Use fiberglass guy-strain insulators rather than the small porcelain guy-strain insulators. Guys are often the lowest flashover path because they are attached right near the top insulator. Ground wiresEliminate unneeded ground wires on poles. Ground wires near the primary can short out significant insulation provided by wood poles and crossarms. Pole-protection assemblies are especially bad and can lead to many induced-voltage flashovers. Fuse cutoutsFuse cutouts are often attached such that a lower-than-normal flashover path is available. Arrange the cutout so that the attachment bracket is mounted on the pole away from any grounds, including guy wires. MetalAvoid using conducting hardware where appropriate. Wood or fiberglass provides significant extra insulation.

The main recommendation for grounding is:

GroundingGood grounding confines a surge to the vicinity of the strike point; if grounding is poor, a voltage surge can propagate further out and subject more equipment to possibly damaging surges. Better grounding also reduces customer-side surges. That said, improving grounding is not necessarily a panacea for all lightning-caused power quality problems. It will not solve all problems or drastically reduce lightning-caused faults. It is a good goal to have good grounding but not worth taking drastic measures to achieve.

For areas needing very high power quality, consider line protection as part of a hi-tech design. While expensive, such designs can eliminate most lightning-caused faults, if they are designed with care. Options include:

Shield wireA shield wire design can achieve high performance if high insulation levels and good grounds (less than 10 ohms preferably) are available. ArrestersArresters can be used on the top phase like a shield wire (requires good grounding and insulation on the unprotected phases) or on all three phases (insulation and grounding are 7-31

Lightning Protection, Grounding, and Arresters

much less important). For significant fault-rate improvement, arresters must be installed on every pole or every other pole.

Shield wire and arrestersUsing both a shield wire and arresters can provide exceptional performance. The shield wire provides a path to absorb most of the lightning energy, and the arresters protect against backflashover.

7-32

8
EQUIPMENT ISSUES
This section covers several miscellaneous issues that can impact reliability and power quality. Both equipment design and equipment material can impact reliability. Age is also an important consideration when considering maintenance or replacement.

Fiberglass Standoffs
Polymer standoffs are a commonly used alternative to the standard wooden cross arm in modern distribution system construction. As the number of units, and the amount of time in the field has increased, several utilities have documented cases in which a fiberglass standoff has unexpectedly flashed over. The phenomenon was initially witnessed by line crews after making repairs subsequent to a storm. When the distribution circuit was re-energized, a standoff assembly flashover was observed. The equipment where the flashover occurred was replaced, and the failed standoff sent for evaluation. Other than severe weathering, and arc damage, the standoff was intact. Since occurrences seem to be rare, there has been very little research into this issue, and the body of knowledge regarding it is therefore rather small. The following discussion presents the most plausible theory to explain the standoff failures based on currently available information237, but other factors that are yet unknown may also play a role in the failures.

237

Crudele, F. D., "Degradation and Flashover Mechanisms of Fiberglass Standoff Brackets," T&D World Expo, Indianapolis, IN, May 25, 2004.

8-1

Equipment Issues

Figure 8-1 Wooden Cross Arm Construction

Historically, distribution line construction utilized ceramic insulators mounted on metal pins and wooden cross arms to support the phase conductors as shown in Figure 8-1. Although wooden cross arms have performed very well, being both strong and durable, they are also heavy and labor intensive to assemble. As an alternative, the insulator industry began producing polymer based replacements for the wooden cross arms in the form of fiberglass reinforced polymer (FRP) standoff brackets. A typical polymer cross arm installation is shown in Figure 8-2. The main advantage of the fiberglass standoff over traditional wooden cross arms lies in the speed and ease of installation resulting in lower labor costs. Therefore, even with a slightly higher materials cost, the fiberglass standoff can still show an economic advantage.

8-2

Equipment Issues

Figure 8-2 Fiberglass Standoff Construction

A polymer standoff consists of a fiberglass reinforced polymer (FRP) core with metal endfittings bonded or clamped to each end. One end-fitting provides a flat base where the standoff can be mounted to the utility pole while the other end-fitting provides a threaded connection point for a pin-type insulator (Figure 8-3). FRP is used for the core because it is relatively inexpensive, easy to fabricate, and mechanically strong. Under dry conditions FRP is nonconductive. The FRP core is coated with a UV resistant paint to give the standoff a weather resistant finish.

8-3

Equipment Issues

Flat End-Fitting for Attachming to Utility Pole

Threaded End-Fitting for Attaching a Pin-Type Insulator Fiberglass Reinforced Polymer Body

Epoxy Paint Worn Away Leaving FRP Core Exposed

Epoxy Paint Still Intact

Figure 8-3 A Weathered and Damaged Fiberglass Standoff Showing the Various Components

Charring Due to Flashover Exposed Glass Fibers

Metal Base Vaproized by Power Arc

Figure 8-4 Damaged Fiberglass Standoff

8-4

Equipment Issues

Vaporized Metal Base

Exposed Glass Fibers

Figure 8-5 Close-up View of Damaged Fiberglass Standoff

Mechanics of the Flash Over The failure of the fiberglass standoff is the result of a combination of factors, each of which by itself would probably not lead to a flashover. The first cause to be considered is the poor weathering characteristics of the epoxy coating on the standoff. The epoxy coating is seen to degrade rapidly when exposed to the UV present in natural sunlight. Figure 8-6 through Figure 8-8 show weathered standoff brackets. These pictures were taken as part of an informal field survey. The location was selected at random and did not have any remarkable features that would accelerate ageing (such as chemical or food processing plants or other unique contaminant sources).

8-5

Equipment Issues

Figure 8-6 Vertically Mounted Fiberglass Standoff Bracket Showing Surface Degradation

Note that the pole top is split, and the opposite standoff is pulling away from the pole Figure 8-7 Horizontally Mounted Fiberglass Standoff Bracket Showing Surface Degradation

8-6

Equipment Issues

Figure 8-8 Fiberglass Standoff Bracket Showing Surface Degradation

As the exterior coating degrades it exposes the underlying fiberglass reinforced polymer core as shown in Figure 8-4 and Figure 8-5. In fact, it is not uncommon for half the length of the core to be exposed after just a few years in the field. When the core is exposed, precipitation falling on the standoff is wicked into the core by the glass fibers thus making sections of the core conductive and reducing the overall creepage distance of the standoff/insulator assembly. This alone should not pose a great threat because the pin-type insulator, which separates the standoff from the conductor, should provide an adequate level of insulation. However, the dielectric relationship between the ceramic pin-type insulator and the fiberglass standoff is such that they form a capacitive voltage divider as shown in Figure 8-9. The divider is characterized by the following equation:
Vstandoff = Vline- to-ground C bell C bell + C standoff
Eq. 8-1

The capacitance of the ceramic insulator is much larger than the capacitance of the fiberglass standoff, thus Vstandoff is nearly equal to Vline-to-ground. This means that most of the line voltage is held-off by the fiberglass standoff and not the ceramic insulator. This does not normally pose a threat under dry conditions. However, as mentioned above, when precipitation is wicked into the fiberglass core it becomes conductive thereby reducing the creepage distance. If the reduced creepage distance is not sufficient to hold off the voltage, and the line potential does not properly shift to the insulator when the standoff becomes conductive, then the standoff flashes over.

8-7

Equipment Issues

Vl g

C bell
C bell V standoff = (V l - g ) C bell + C standoff

C standoff

Figure 8-9 Capacitive Voltage Divider Created Between a Ceramic Pin-Type Insulator and the Fiberglass Standoff

Are Flashovers More Likely After Thunderstorms? The common impression within the utility industry is that the fiberglass standoffs are more likely to flashover shortly after thunderstorms. However, there is no specific evidence to either prove or disprove this. It is possible that the standoff failures appear to be more frequent following thunderstorms because line crews are more frequently in position to observe the failures during this time period since they are out repairing storm damage. Nevertheless, there are some interesting characteristics of the post-storm time period that lend credibility to the idea that the standoff flashovers are more common in the post storm timeframe. There are two factors that may cause increased flashover activity after thunderstorms. The first deals with the wetting and drying characteristics of the standoff and the impact on its conduction properties. The second deals with high-frequency transients which may occur when reenergizing a line after a storm related outage or line repair. During the storm, precipitation is wicked into the core through areas of exposed fiberglass. The wetted core provides a conductive path causing the line potential to shift from the standoff to the insulator. Once the storm passes and the standoff begins to dry, dry bands begin to form, which disrupt the conductive path and cause the potential to shift back to the standoff. If the potential across the standoff is great enough an arc will form over the surface to bridge these high resistance dry bands. This dry band arcing process leads to carbonization of the polymer creating permanently damaged conducting sites that can grow over the surface of the rod. This process can then lengthen to eventually short the entire standoff resulting in a flashover. Coincidentally, the time frame in which the candlestick becomes partially dry is also when line crews are likely to be re-energizing circuits that had tripped due to storm activity. When a circuit is re-energized a transient is created and the line can experience an increased voltage and a high frequency ringing on top of the 60-Hz power frequency. As if the overvoltage wasnt bad enough, the voltage divider between the ceramic insulator and the fiberglass standoff is also more likely to be dominated by their relative capacitances, as opposed to resistances, under the higher frequency conditions thus furthering the extent to which the line voltage must be held off by the fiberglass standoff. 8-8

Equipment Issues

Laboratory Demonstration A laboratory experiment was set up to investigate the discharge behavior of a fiberglass standoff upon energization after the passage of a storm. In this test an I6, complete with a tied conductor, was placed on top of a standoff. The standoff used had previously been in service for sufficient time for most of the painted surface to have been lost due to weathering, revealing much of the underlying resin and reinforcing glass fibers. A small amount of distilled water was sprayed over the entire assembly to simulate wetting due to passage of the storm. The output of a voltage source was connected to the conductor on top of the I6, while the bottom of the standoff was grounded. The applied voltage was chosen to be at a high frequency to correspond to the oscillations present in a typical energizing switching operation. Numerous discharges were found to appear along the length of the standoff, occasionally accompanied by discharges on the I6 and conductor assembly. Figure 8-10 shows an example of the discharge behavior.

Figure 8-10 Discharges Over the Surface of an Energized Fiberglass Standoff

The voltage applied in this test was 35 kV, at a frequency of 1 kHz. In Figure 8-10 it can be seen that there are several coincident arcs along the length of the fiberglass rod, indicating that most of the applied voltage lies across its surface rather than on the insulator that it is supporting. Degradation of Other Fiberglass Apparatus Other fiberglass reinforced polymer (FRP) products often experience the same type of field degradation whether or not they are under electrical stress. For example, fiberglass guy strain insulators which are used to replace ceramic guy insulators (commonly referred to as johnny balls) often exhibit similar surface degradation due to UV exposure and environmental contaminants. They are afforded somewhat better UV resistance compared to the insulator standoffs because they are veiled rather than painted but they still show similar degradation over time. The veiling technique is used to provide the guy strain insulators with better abrasion 8-9

Equipment Issues

resistance. Although they are not normally under electrical stress, it is unclear how the surface degradation affects the mechanical strength of the insulators. Possible Solutions There are several possible solutions that can be applied to avoid the standoff flashovers:

Different insulators It may be possible to use insulators with a lower capacitance thus altering the capacitive divide ratio and placing more of the working voltage across the insulator instead of the standoff. Some polymer insulators may offer a lower capacitance than the ceramic insulators. However, capacitance is highly dependent on the insulator shape, so long, skinny insulators, much like the candlestick, would most likely offer the best chance for a more favorable divider ratio. Different or additional sheath material The problem may be remedied by keeping the fiberglass core from being exposed. One possibility would be to jacket the core in several millimeters of silicone rubber. While it would add to the cost of the unit, it would also provide far superior weatherability and keep the inner core from being exposed. Other materials such as EPDM or an improved paint formulation could also be employed. Conductive standoffs By incorporating conductive materials in the core of the standoff, or even by using a metallic standoff, the voltage distribution would favor a condition where the line voltage would be across the insulator only, at all times. This approach achieves the goals of fast and inexpensive construction while reducing the flashover potential of the standoff / bracket. Wood construction While the fiberglass standoffs are lighter and less costly to install than the traditional wood cross arms, which is why they are used, the wood cross arm does not pose the same flashover threat as the fiberglass standoff. While reverting back to wood construction is a possibility, it is also the least attractive solution because it represents a step backwards in terms of cost and ease of installation.

The Need for Further Study Little information is known within the utility industry regarding the standoff failures. One reason for this is that the failures are seldom observed and hard to diagnose. The failures do not always cause a protective device to operate and when they do the standoff may function normally once the line protection clears the fault and the circuit is re-energized. One consideration that remains relatively obscure is what exactly the electrical path is when the standoff flashes over. It is not known whether it presents a line-to-ground fault, line-to-line fault, or both. The main source of information on the occurrence of these failures has come from line crews who witnessed the failures in the field. There has been very little laboratory research done to investigate the cause of these failures and possible remedies. It would be beneficial to the industry for more formal research to be untaken to study the processes that lead to such failures.

8-10

Equipment Issues

Polymer Insulators
While they have not yet gained widespread acceptance in the utility industry, polymer based insulators are being used to a greater extent than ever before in the construction and repair of 238 electrical distribution lines. The first of the modern polymer insulators emerged in the 1960s . From that point onward, manufacturers have constantly strived to increase their products performance while decreasing costs. Over time, the forward progress in technology has resulted in many changes to the physical design of the insulators as well as changes in the base materials and fillers they are made from. Unfortunately, the constantly evolving nature of the polymer or non-ceramic insulator (NCI), coupled with its relatively recent development, has made it difficult to obtain long-term field performance data. For this reason the long-term reliability of polymerbased insulators is still somewhat unknown. In order to help fill in this gray area, a great deal of work has been done in accelerated ageing and laboratory testing of polymer insulators. There are however many ageing and stress variables to consider including environmental considerations, variations in materials, and physical differences in the designs from different manufacturers. For these reasons it is still not possible to definitively determine whether polymer based insulators will exhibit an operational life comparable to the ceramic insulators they are intended to replace. However, there is a strong case that polymer insulation can outperform ceramic over relatively short service intervals, especially under heavily polluted conditions. For more background on polymer insulators, see Hackam239, Mackevich and Shah240, Mackevich and Simmons241, and Simmons et al. [1997]. Composition of Polymer Insulators Modern polymer insulators are molded from one of four main compounds: ethylene-propylenediene-monomer (EPDM), silicone rubber, high-density polyethylene (HDPE), or another thermoplastic material. In addition to these base polymers, the insulators contain a variety of fillers as well as other agents to enhance their performance. In fact, the compounds for modern non-ceramic insulators are rather complicated and usually consist of 10 or more different ingredients. Most of the materials in the formulation fall into the following categories:

Elastomers This is the base polymer, usually EPDM, HDPE, silicone rubber, or thermoplastic material. Vulcanizing agents These agents cause a chemical reaction that results in cross-linking of the elastomer material. Cross-linking makes the material stiffer and more temperature stable. Coagents The coagents protect the already formed cross-link bonds and help to promote new bonds.

Hall, J. F., "History and bibliography of polymeric insulators for outdoor applications," IEEE Transactions on Power Delivery, vol. 8, no. 1, pp. 376-385, 1993. 239 Hackam, R., "Outdoor HV composite polymeric insulators," IEEE Transactions on Dielectrics and Electrical Insulation, vol. 6, no. 5, pp. 557-585, 1999. 240 Mackevich, J. and Shah, M., "Polymer outdoor insulating materials. Part I: Comparison of porcelain and polymer electrical insulation," IEEE Electrical Insulation Magazine, vol. 13, no. 3, pp. 5-12, 1997. 241 Mackevich, J. and Simmons, S., "Polymer outdoor insulating materials. II. Material considerations," IEEE Electrical Insulation Magazine, vol. 13, no. 4, pp. 10-16, 1997.

238

8-11

Equipment Issues

Antidegradants These retard the deterioration of the rubber compounds. Rubber compounds are degraded by oxygen, ozone, heat, and UV light and therefore the antidegradants are also called antioxidants, antiozanants and inhibitors. Processing aids The processing aids help in the flow, mold, and release stages of the manufacturing process. Fillers Fillers are used to reinforce physical properties, impart processing characteristics, and reduce costs. Two common fillers are alumina trihydrate (ATH) and silica. It is typical to include 10-20% silica for rheological control and ~50% ATH to act as a flame retardant. Adding ATH also provides a cost savings, since ATH is generally less expensive than the polymer base material. Coupling agents The coupling agents aid in bonding the filler and elastomer materials. Plasticizers and softeners These are used to adjust the flow and flexibility of the material during the molding process and in the final product. Special purpose materials These are typically the manufacturers secret ingredients that are used to impart certain characteristics to the material that help set it apart from the competition.

The exact formulation of the polymer compounds are proprietary to each manufacturer and are closely guarded trade secrets. It is also common for manufacturers to alter their formulation every few years as they work to improve their products and reduce costs. Depending on the particular application, a polymer insulator may be molded as one piece which is composed entirely of the polymer material or it may be made from several pieces, not all of which are polymer based. For example, pin-type polymer units that are intended to replace ceramic bell insulators are injection molded as a one piece unit composed entirely of polymer material and resemble the traditional ceramic bell in shape (Figure 8-11). One the other hand, strain-type polymer insulators utilize a polymer jacket and weather sheds molded to a fiberglass reinforced polymer core with metal end-fittings clamped to each end as shown in Figure 8-11. Available Designs Polymer insulators are available for just about any distribution application that was traditionally performed by a ceramic insulator including:

Pin-Type Insulators Strain-Type Insulators (dead-ends) Post Insulators Standoffs (cross arm replacements) Guy Strain Insulators

Figure 8-11 shows some of the different types of polymer insulators available. Polymer insulation was initially adopted on a wider scale by the transmission industry because of the tremendous weight savings offered by polymers. This was particularly beneficial when 8-12

Equipment Issues

upgrading transmission voltages on towers that could not handle the weight increase that additional ceramic insulation would have caused.

Figure 8-11 Different Polymer Insulator Designs

Advantages of Polymer Insulation Polymer insulation has several benefits when compared to traditional ceramic insulation, namely:

Weight reduction Polymer based insulators are much lighter than their ceramic counterparts. The average 15-kV class strain-type polymer insulator weighs approximately 2.5 pounds while a 7.5-kV class strain-type ceramic insulator averages 5.5 pounds. Furthermore, a single 15-kV polymer insulator can be used where two 7.5-kV ceramic insulators were needed previously. This corresponds to an approximate 75% reduction in weight! Flexibility and impact resistance Polymer based insulators offer a much higher level of impact resistance compared to ceramic insulators. Since the polymer material is less brittle than the ceramic material, the polymers are less likely to be broken during transportation and installation. Polymer insulators are fairly tolerant to being dropped and offer great advantages in resistance to vandalism. While a gunshot will normally shatter a ceramic insulator, polymer insulators have a tendency to allow the bullet to pass through the insulator sheds leaving behind a hole but remaining for the most part in tact.

Drawbacks of Polymer Insulation Utilities, which are historically conservative in nature, have been slow to adopt polymer insulators into their construction practices. Switching over to polymer insulators presents some new ground for many utilities who are primarily concerned with the following drawbacks: 8-13

Equipment Issues

Lack of long-term field performance data Since the modern polymer insulator industry is relatively new, especially given the expected length of service of distribution insulators, there is little long-term field performance data available. Another factor complicating the collection of long-term data arises from the continually evolving nature of polymer insulators. Since manufacturers alter the composition of their products every few years many products that have undergone field and laboratory testing are no longer available. Susceptibility to tracking and erosion Polymer insulators can be more susceptible to tracking and erosion than ceramic insulators. The polymer material is particularly vulnerable to corona cutting in which corona discharges on the surface of the insulator cut channels into the polymer material. Corona cutting leaves behind carbon tracks which can act to short the insulator or the corona cutting can degrade the insulator to a point at which is fails mechanically.

Degradation of Polymer Insulators Like ceramic insulators, the performance of polymer or non-ceramic insulators is degraded by 242,243 . Insulator degradation usually begins as exposure to the natural environment in the field surface erosion, which results in reduced surface hydrophobicity and may eventually culminate with dry-band arcing and insulator failure. Exposure to ultraviolet light and ozone, both of which are naturally occurring and created by corona discharges, degrades polymer materials. Outdoor insulation is also constantly exposed to airborne pollutants from a variety of sources. A few of the major contaminants are road salt that is levitated up from the roadway, sea salt, cement dust from cement plants, and bird droppings. Additionally, hematite (Fe2O3) is found on insulators in strong concentrations near steel plants; and anhydrous gypsum, silica, and bassanite are found on insulators in both agricultural and industrial areas. These pollutants settle out of the air onto the surface of the insulator coating it in a layer of pollution dust. Furthermore, the acids found in acid rain also attack the insulator surface causing erosion of the polymer material. Surface degradation results in reduced surface hydrophobicity. The roughened surface with reduced hydrophobicity is more likely to hold contaminants and water thus forming a conductive film on the insulators surface. The actual conductivity of the film depends on the surface contamination as well as the conductivity or the precipitation. Once contaminated, leakage current will flow through the film layer creating heat from the process of Joule Heating. The heat created causes non-uniform evaporation of the water on the insulator surface leading to dry-bands which are surrounded by wet areas. The dry-bands present an area of high resistance in the conductive path on the surface of the partially wetted insulator. This process effectively shortens the creepage distance of the insulator and allows for a high voltage drop over the small distance of the dry-band. The increase in electric field across the dry-band is enough to cause small discharges or scintillations. The discharges will carry more current as they grow in size and intensity. The arcing discharges produce a large amount of localized heat, which decomposes the polymer and allows the free carbon to rise to the surface. This process forms a conductive, carbonized path, or track, that will eventually short the electrodes of the insulator.
Crudele, F. D., "Accelerated Ageing of Outdoor Polymer Insulating Materials Using Tracking Wheel Techniques," Master of Science Thesis, Clarkson University, Potsdam, NY, 2002. 243 McGrath, P. B., Crudele, F. D., and Burns, C. W., "Accelerated Aging of Polymers Using Synthetic Acid Rain Solutions," Conference on Electrical Insulation and Dielectric Phenomena, Kitchener, Ontario, Canada, October 14, 2001.
242

8-14

Equipment Issues

Polymer insulators may also suffer from corona discharges. The UV generated by the corona discharge promotes surface erosion and the heat generated by severe corona discharges can cut into the polymer surface resulting in what is known as corona cutting. Corona cutting is rare at the distribution level so loss of hydrophobicity from surface erosion tends to be the greater concern. Issues Not Related to Electrical Performance Other issues that are not based on insulator performance can influence utilities to apply polymer based insulation. One major factor is that polymer insulators have reached a cost at or below that of ceramic insulators. In fact in many utilities, the purchasing department is driving the switch to polymer insulators. A second driver in the switch to polymer insulation is a continually diminishing supply of ceramic insulators. Several ceramic insulator manufacturers have stopped production or gone out of business in the past several years creating some concern about their future availability. Current Usage Trends Non-ceramic insulators currently represent the majority of newly installed high voltage insulators in North America. Furthermore, some estimates indicated that as much as 70% of new 244 insulator installations are non-ceramic . Early insulator formulations left much to be desired and attached a stigma that polymer insulators are just beginning to overcome within many utilities. In general, utilities appear to look favorably upon the performance of their existing NCI installations, especially those utilizing the latest product offerings. The main source of concern regarding polymer insulation continues to stem from their uncertain service life. Modern polymer insulators have not exhibited reduced service life, they just havent been in the field long enough to prove that they are capable of 30 years or more of reliable service. Overall, it seems that the migration to non-ceramic insulators is in full swing and may almost be a necessity as the supply of ceramic insulators continues to dwindle.

Stress Cracking of Fuse Cutouts

Some utilities have reported that a portion of their porcelain fuse cutouts are experiencing stress cracking of the insulator that can lead to it breaking apart. During fuse insertion/removal, the crack can lead to a break that results in faults or swinging energized parts hanging from or near the cutout. A weakened cutout can also break on its own (unattended) due to external loads or during the forces associated with a fault clearing operation of the fuse. Cracks in porcelain can allow water ingress over time into the material itself, which may eventually lead to flashovers within and around the porcelain. These failures can pose a risk to crews working on circuits as well as the public, and they can degrade power quality and reliability. The nature and extent of this problem is currently under study within the industry. For the most part, the problem appears to be limited to one type of porcelain fuse cutout (manufactured mainly 10+ years ago).

Hackam, R., "Outdoor HV composite polymeric insulators," IEEE Transactions on Dielectrics and Electrical Insulation, vol. 6, no. 5, pp. 557-585, 1999.

244

8-15

Equipment Issues

However, the industry needs to collect more data on this to determine if the issue is more widespread and is significant enough to warrant industry wide attention. Porcelain Cutout Construction and Crack Formation Porcelain insulated fuse cutouts use three metal pins cemented with a grouting compound into holes in the porcelain insulator (these are often called potted pin cutout designs). These metal pins are located at the top and bottom of the insulator to hold the source and load conductor connecting hardware, and a third pin is at the center body of the insulator to attach the cutout mounting bracket (see Figure 8-12). The stress cracking that has been observed in some cutout units is believed to originate from within the porcelain at the locations where these pins are cemented into the insulator. At these locations, the porcelain is subjected to stresses due to the electrical, mechanical and environmental operating conditions to which the cutout is exposed. Over time, water entry into cement material can occur. Freeze/thaw cycles of this water result in stress cracking of the porcelain at the cement/porcelain interface. Besides water entry, other processes, such as the differential expansion between the porcelain and metal, may also create some of the stress cracks.
From Source

Porcelain Insulator

Utility Pole Locations Where Metal Pins are Cemented In Place and Cracking May Be Initiated Fuse Element in Holder

To Load

Figure 8-12 Typical Porcelain Fuse Cutout Configuration Showing Cemented Metal Pins in Porcelain

Once a crack is initiated, it can grow through repeated stress cycles until mechanical cutout failure occurs. A cutout can remain partially cracked for some time without apparent malfunction and would even require very close physical inspection to see the cracks in some cases. The force of a lineman inserting or removing a fuse can easily be more than enough to completely break a 8-16

Equipment Issues

weakened (cracked) unit. A breakage at the moment of fuse insertion/removal could easily result in the lineman accidentally causing a fault so this becomes a safety concern for the utility. Approaches to Risk Management for Porcelain Cutout Cracking British Columbia Hydro (BC Hydro) is one utility that has studied the porcelain cracking issue and developed a strategy for managing the risk. They published their results in a recent technical paper245. The BC Hydro experience shows that they have approximately 308,000 porcelaininsulated fuse cutouts on their distribution system and most of these are in good condition however, a small portion have begun to crack. To mitigate the risk of defective cutouts and to maintain reasonable reliability of their distribution system, BC Hydro adopted a Risk-Based Least Cost Strategy for managing risk associated with their porcelain fuse cutouts. The strategy was developed by Powertech Labs and included the following steps:

A random sampling process of the existing cutout population was performed to determine the portion of the fuse cutout population likely to be subject to a cracking condition. The reliability of cracked cutouts under the range of typical operating conditions was assessed. The risk that cutout failure poses to line maintenance personnel and the public was assessed and remedial actions evaluated to reduce risk. Cost-benefit analysis of the most appropriate remedial actions was assessed.

The random sampling program evaluated 13,529 fuse cutouts across 52 districts in the BC Hydro service territory. Based on the sampling results, the percent of the installed population with cracking ranged from 0.27% in the district with the lowest number of defective units to 8.0% in the district with the highest number of defective units. For the entire population (all districts), the percent of the population with cracking was calculated to be about 0.84% based on the random sampling. No cracking was observed in units purchased after 1997. BC Hydro indicated that this was due to more stringent quality control included in the purchase specification. A fault tree was developed for BC Hydro that considered the possible risk outcomes associated with cracked cutouts. The analysis considered the probability of failure when crews were attempting to operate the cracked cutout as well as the probability of failure when crews are simply working nearby (climbing the pole and causing vibrations). Based on historical injury data, laboratory tests of the forces needed to fail a weakened cutout, and engineering judgment, the probability of injury associated with a lineman attempting to work with or near a partially cracked (weakened) cutout was conservatively estimated to be 1 in 1000 per work operation (0.1%). The probability of loss of life was estimated to be 1 in 5000 per work operation (0.02%). To address the fuse cutout problems, BC Hydro considered a remedial action plan that offers the lowest net present value (NPV) cost over a 10-year period. The cost model included factors such as liability, outages, materials and labor. Three remedial approaches were considered at the district level:
Li, H., Bhuyan, G., and Tarampi, D., "Risk-Based Least-Cost Strategy for Managing Risk Associated with Porcelain Cutout Asset in BC Hydro System," 8th International Conference on Probabilistic Methods Applied to Power Systems, Iowa State University, Ames Iowa, September 12-16 2004.
245

8-17

Equipment Issues

1. No actionSimply replace cracked cutouts as they are found in the field this was the existing standard practice. 2. Selective replacementReplace fuse cutouts, regardless of their being cracked or uncracked, when crews are working at cutout pole. Also, replace in high-risk areas. 3. Blanket replacementInitiate program to rapidly replace all pre-1997 porcelain cutouts throughout the system. Based on the NPV analysis, option 3 was determined to be best (lowest NPV cost) for two of 52 districts, option 2 was best for 18 districts, and option 1 was best for the remaining districts. It should be recognized that while the BC Hydro work focused heavily on the safety of crews and the public, that there also is a reliability benefit of applying the proposed remedial plans since this would reduce faults and accelerate the speed of power restoration efforts by crews. Several other Canadian utilities besides BC Hydro have had an interest in this issue. As a result, CEA Technologies, Incorporated (CEATI), a Canadian research organization, has prepared a report titled Condition Assessment of Porcelain Cutouts (#T044700-5031) that evaluated methods for detecting cracked cutouts in the field. The report also contains some data on cutout failures. In the US, some utilities have also expressed interest in the issue and Georgia Tech's National Electric Energy Testing Research and Applications Center (NEETRAC) has recently initiated research to assess the scope of the problem. The project involves laboratory tests and analysis on failed units and was just beginning as of early 2005. Results of the NEETRAC investigation may be available later this year (2005) or next year. To obtain a better understanding of the scope of the cutout cracking issue, many North American utilities, a manufacturer, and industry experts were informally surveyed about their experiences with cracking fuse cutouts. The consensus of the discussions was that most utilities are not having a problem at this time but that roughly 1/5th (20 percent) of those responding reported a problem. Those that have a problem reported it was primarily with one particular brand made nearly a decade ago. 15- and 25-kV fuse cutouts were the most commonly cited voltage ratings where the problem occurred. There is more concern about this issue in colder climates (such as at Canadian Utilities and northern US utilities) which makes sense given how the freeze/thaw cycles likely impact the formation of cracks. Based on the discussions with industry experts, manufacturers and utility engineers, the susceptibility of the cutout to this kind of cracking is greatly a function of the manufacturing quality control and design, and it can be virtually eliminated with proper manufacturing processes, materials, and sealing methods. This statement tends to be confirmed by the fact that BC Hydro was able to essentially eliminate cracking in units purchased after 1997 by specifying specific manufacturing characteristics and quality. As a result, it is felt that the cracking problem has nothing to do with the general viability of porcelain as an insulation medium; in fact, porcelain has shown excellent long-term aging characteristics in its history of use. Instead, the problem is with the manufacturing process and materials used in the design of the particular cutouts in question. Some engineers feel that the industry movement from the old banded fuse cutout design of several decades ago, to the more modern potted pin based designs, while lower in cost, not only results in a potential water ingress point if the potting material cracks, but also results in 8-18

Equipment Issues

increased stress on the porcelain. The potted pin approach subjects the insulator to the weaker tension mode of strength rather than in its stronger compression mode as the banded designs did. These may be reasons for the increased cracking happening in some potted pin porcelain cutouts. These characteristics for potted pin designs put increased importance on having a good manufacturing design and process control compared to the earlier banded designs. An area that needs to be looked at is the reporting of cutout failures. It is hypothesized that some cracking problems may be underreported because they are reported simply as cutout flashovers or faults in the utility outage reporting database and this does not stand out as a cracking issue in the minds of utility engineers that see this data. These may be instead listed as a lightning flashover, contamination flashover, or some other type of non-cracking related failure. However, some of these faults may have originated from water ingress into stress cracks which led to reduced voltage withstand strength, tracking and eventual failure (flashover). If this is the case, then a larger portion of utilities could be experiencing the cracking issue than is accounted for in the informal survey discussed above. The data that has been published in the industry technical literature so far is insufficient to determine if the cracking problem is widespread or simply inconsequential and not worthy of much attention. The pending results of the NEETRAC investigation, possible future EPRI studies, and other studies at utilities should eventually provide the data needed to address this issue. For now the main advice to utilities when considering this issue is to make sure that the cutout manufacturer from whom they purchase equipment recognizes the potential problem and has implemented suitable safeguards in the design and manufacture to limit moisture ingress into the pin locations and limit stresses that could cause cracking. Some utilities have decided to change over to polymer insulated fuse cutouts as a solution to cracking, but this approach, in the long run, may not lead to any higher reliability since overall porcelain units have generally performed well in the field and the porcelain insulator surface has well established and excellent aging characteristics. Polymer insulators, being newer, do not yet have as much long-term aging experience. Remedial actions to detect cracking in the field and change out any existing units that are cracking must be done on a cost effective basis similar to that described in the BC Hydro paper. However, the CEATI work indicated that it can often be difficult to detect the small cracks visually in the field so how this process is implemented practically without removal of the cutout is not clear. Some cracks can be easily detected, but others may go unnoticed.

8-19

9
CUSTOM CONFIGURATIONS
There are many customized ways that utilities can improve reliability and power quality. Most custom solutions targeted at specific customers or groups of customers.

Reactors
The use of line reactors can improve the voltage levels on distribution circuits during line-to-ground faults on adjacent circuits. Line reactors are particularly useful in areas with high available fault current. Neutral reactors on the substation transformer can also provide improvement under the right conditions. Benefits

Improves voltage level of distribution circuit during faults on adjacent feeders. Reduces fault availability, allowing for protective devices with lower fault-duty ratings.

Negatives The use of line reactors reduces fault availability, which can make coordination of protective devices difficult if the fault availability is comparable to load current. Economics Line reactors require a certain amount of valuable station real estate. The cost of the reactors and the associated real estate can be offset in some cases by allowing the use of protective devices with lower interrupt ratings. If a sensitive customer is being adversely affected by faults on adjacent feeders, the application of line reactors may be a cost-effective solution. While still a cost, neutral reactors are less expensive than line reactors and take up less space. Bottom Line Line reactors can improve the voltage levels on distribution circuits during faults on adjacent circuits. Line reactors also reduce the amount of available fault current on the circuit. The reduction in fault current can be helpful in areas with high levels of fault current, such as stations with large or parallel transformers. In the case of circuits with low fault currents, line reactors can further complicate the coordination of distribution-protection devices. Because of the cost

9-1

Custom Configurations

and engineering, reactors are best used to help specific sensitive customers. Neutral reactors can provide improvement, especially with delta-wye customer transformers. Introduction Current-limiting reactors are traditionally considered as a means of limiting fault duty for circuit breakers where large step-down transformers are connected to the transmission system. Short-circuit duties for line breakers may exceed their capability unless intentional impedance is placed in the path between the source and the fault location. The effects of current-limiting reactors on power quality are not generally considered when their use is indicated on the basis of increased fault duty. Their use in reducing fault duty will be examined as well as their potential use for improving power quality events resulting from distribution line faults. Available Short-Circuit Current: Traditional Substation Design In the most fundamental design of distribution stations, a single transformer supplies a bus with a limited number of connected distribution circuits. The fault current for a line fault is supplied by the sole transformers and is equal for both the transformer low-side breaker and the line circuit breaker (see Figure 9-1).

Figure 9-1 Fault Duty on Single Transformer for Line Fault

As an area load continues to grow, the demand is frequently met by the addition of a second step-down transformer and distribution bus. Figure 9-2 shows the effect of a second transformer on fault-current availability. In this case, the fault current on the line circuit breaker is greater than on either of the transformers low-side breakers.

9-2

Custom Configurations

Figure 9-2 Fault Duty for Two Transformers for a Line Fault

Utilities serving dense load concentrations in urban areas often install triplex stations with individual transformer ratings of 65 MVA at 13 kV and 150 MVA at 34 kV. These sizes clearly exceed the capacity of conventional circuit breakers and result in excessive fault duties for customer equipment located downstream from the feeder breaker. Drop-In Substations As an alternative to increasing capacity at existing stations or in areas where only spot system relief is required, the rather recent use of prepackaged plunker-type stations has become popular. These stations are designed to be installed by tapping an existing transmission line without the requirement for additional right of way. To minimize both size and expense, these drop-in stations are necessarily of limited power capacity (in MVA). As a result, the exiting distribution circuit may have a reduced continuous-current rating and an accompanying limited interrupting capability. Although the transformer power rating is relatively small, design for fault clearing often dictates a higher-than-normal transformer impedance as a limiting mechanism. A higher-than-normal transformer impedance will be shown to provide the worst possible voltage-sag performance of the designs considered. Distribution System Sag Performance Even if the time and specific location of distribution system faults are unpredictable, their occurrence is guaranteed, and frequency is well established for various configurations. As documented in many power quality studies, the most frequent distribution event is a sag, not an outage. The ratio of sag-to-outage events is primarily a function of the number of distribution lines originating at the same station bus. For every fault on the distribution system, a sag is experienced on the source bus, and every customer fed from that bus experiences a sag. Consider a bus that feeds six identical lines, one of which is faulted. For every customer fed from the 9-3

Custom Configurations

faulted line, five will experience a sag. Data from previous distribution power quality studies246 have been superimposed upon the traditional equipment sensitivity curve (see Figure 9-3).

ITI CURVE AND PQ DATA . . .

120% 100%

Objective

80%
Percent of Nominal RMS

60%

40%

20%

0% <1c 2c 4c
Duration in Cycles (c) and Seconds

6c

20c

1s

5s

30s

120s

LOWER UNDERVOLTAGE BOUNDARY BY THE DIFFERENCE IN BUS VOLTAGE WITH LINE REACTOR

Figure 9-3 Data from EPRI Distribution Power Quality Study Overlaid on the ITIC (Information Technology Industry Council) Curve

Each point in Figure 9-3 indicates an event as seen by the aggregate of customers fed from a typical distribution system. A representative equipment-sensitivity curve has been overlaid on the event points. Points in the Objective band are of little interest because connected equipment should not be adversely affected in this region. The stepped area under the Objective bandbelow the ITIC curveindicates events that are likely to affect customer equipment. Events in the area from the 0% on voltage axis up to the lower limit of the stepped area may disrupt unprotected equipment. Nature of Distribution System Events The basis for an understanding of utility sags lies in three areas:

246

How many events occur (implicitly, how many affect connected equipment)?

An Assessment of Distribution System Power Quality: Volumes 1-3, EPRI, Palo Alto, CA: 1996. TR-106294-V1, TR-106294-V2, and TR106294-V3.

9-4

Custom Configurations

How long does the sag event persist? How severe is the magnitude of the event?

The number of faults on a specific line is in large part a function of line construction, geographic area of service, tree- and animal-related events, and weather-related phenomena. The number of lines fed from the same bus basically determines the number of resultant sags. As seen later, this treats parallel busses as one. A normally closed tie breaker has no effect on sag performance. A closer examination of the event scatter plot in Figure 9-3 will add clarification. How long does the sag persist? All distribution customers will see a fault occurring on the transmission system. Transmission line fault clearing is as fast as possible with no intentional relay delay in the primary relaying of the faulted segment. Breaker clearing times are exceptionally fast when compared to traditional distribution equipment. The scatter events appearing between a 2-cycle and 6-cycle duration may be considered to be either transmission events or extremely close-in distribution faults. The clustering of events below 10% retained voltage and lasting from 20 cycles to 2 minutes may be considered to be faults on the line where the specific monitors were located. The duration would indicate that an instantaneous relay operation did not occur, and depth would indicate an extremely close-in fault. The duration of the scatter event can be correlated to system relaying and fusing practices. The depth or severity of the events is related to the source impedance, available fault current, and distance from the bus to the actual fault location. The lower the source impedance, the higher the available fault current; the further the fault from the bus, the less severe will be the voltage sags on the adjacent lines. Not only is the bus voltage less severely depressed during a fault, but the fault duration may also be reduced by a protection scheme clearing faster. Excessive Fault Duties From the above statement, it would seem intuitive that sag performance would improve by providing as high an available short-circuit duty for the bus as possible. Interrupting duties for both utility line breakers and installed customer equipment impose a limitation on this value. To limit fault duty, the utility may insert an impedance between the sources and potential fault locations or may choose to limit the fault current by opening tie breakers. The location for inserting impedance to limit fault duty may be equally effective at a number of electrical locations. The effect on sag-related power quality performance, however, may be vastly different. Reactor Placement Options As stated earlier, the fault duty of a line circuit breaker is determined by the number of parallel transformers feeding the bus. Two transformers will result in twice the fault current. The installation of a tie reactor immediately addresses two concerns:

Excessive fault duty Continuous losses associated with normal current flow through the line reactor (see Figure 9-4) 9-5

Custom Configurations

If both busses connected via the tie reactor are loaded equally, the normal condition will see no current flow through the tie reactor and thus no losses. The reliability of the bus source is not compromised because the tie is still operated closed and able to accept a loss of transformer (or transmission section if properly configured). The design of the tie reactor should consider not only fault limitation but also voltage effects when all of the circuits are fed from a single transformer.

TIE REACTOR

Figure 9-4 Fault Duty With Tie Reactor

This placement of the tie reactor reduces the fault current for all line faults and has mixed effects on overall station sag performance. Faults on the left bus will result in a lower retained bus voltage during clearing, while the voltage on the right bus (now somewhat isolated by reactance) will be less depressed. Another frequent placement for fault-limiting reactors is the transformer secondary (see Figure 9-5). To maintain current sharing between transformers, a reactor is placed at both low-side positions. This placement effectively limits line-fault current and has the advantage over the tie reactor in that fault duty is reduced on the transformer breaker as well. Disadvantages include reduced voltage regulation as loading changes and continuous losses associated with all connected loads because of the voltage drop in the reactor.

9-6

Custom Configurations

Figure 9-5 Effect of Transformer Secondary Reactors on Fault Duty

The effect of a line fault in this configuration (see Figure 9-5) will be the same regardless of the source bus off which it occurs. Because of the addition of transformer reactors, the short-circuit. duty will be limited as desired. The side effect will be that all line faults are now more severe than before the installation of additional series impedance. The final placement of reactors to be considered will be in the line positions themselves (see Figure 9-6). This option is the most expensive because a reactor set is required for all line sections requiring reduced fault duty. In addition, losses will always be present with the lines in service.

9-7

Custom Configurations

Figure 9-6 Effect of Feeder Reactors on Fault Duty

This reactor placement in Figure 9-6 can dramatically improve sag performance at the bus. That is, the higher bus voltage will be retained during voltage sags for all customers not fed from the line experiencing the fault. The voltage on the faulted line will be depressed even more severely than in the condition without a reactor. However, the resulting voltage sag will likely not be as significant as the fault because the fault will result in unacceptable voltage and may be accompanied by an interruption because of fault clearing. Typical values for reactor impedance are generally 1 to 1 ohms. Different installation details have been employed for reactor placement. Where metal-clad switchgear is used, it is generally necessary to place the reactor beyond the electrical location of the breaker. While effective in limiting fault duty for events in the street, it is not a factor for faults between the breaker and the reactor location. This placement is an accepted practice by a number of utilities. The consequences of a fault in this unprotected area provide a vivid demonstration of the forces associated with high-fault-duty short circuits. For outdoor station design, the option for placement is somewhat more flexible. Reactors in outdoor design are generally installed on bus supports and may be placed between the bus and the breaker or following the breaker but before the street tap.

9-8

Custom Configurations

Power Quality Ramification Research has addressed the reduction in the number of sags as seen by critical customers by the reduction of line exposure by operating the utility system with the tie open. Operating with the tie open certainly eliminates almost all sags that occur for line faults on the bus on the opposite side of the open tie reactor. As shown, the bus voltages associated with sags on the same bus during line clearing are worse. Additionally, this option is not viable for certain transmission configurations. Figure 9-7 shows a common station arrangement.

Figure 9-7 Common Distribution Station Arrangement

In this common configuration, the transmission line and transformer protection zones overlap. A fault above the transformer low-side breaker, including the transmission line to the right, will open both the high-voltage tie breaker and the transformers low-side breaker. Operation with the tie breaker open will cause the loss of the associated distribution bus until the tie breaker closes. Although utility philosophies vary, the tradition practice is to isolate failed components and not rely upon a positive response to re-energize sources. If the tie breaker were operated normally closed, the transformer to the left in Figure 9-7 would usually be thermally capable of carrying the entire station load while the system is reconfigured. The open tie breaker operation was one of many options considered during a recent assessment of utility sag performance where critical customers were located on a number of lines originating from each bus section. Although breaker fault duties were not a consideration, the use of line reactors to mitigate sags because of faults on the adjacent line were found to be a viable option.

9-9

Custom Configurations

Study Findings The station under investigation was operated with two transformers feeding separate 23-kV busses with a pair of normally closed tie breakers. Although only two lines are shown in Figure 9-7 from each bus, each bus fed five or six distribution line sections. The effect on sag performance by the addition of line and tie reactors was determined by the use of a computer short-circuit study performed with and without reactors placed at both the tie and line positions. The results of the fault study are displayed as the retained bus voltage for distribution faults located at various distances from the bus (see Figure 9-8). The as-found condition (with no reactors installed) shows what can be expecteda fault at the station fence (or within the station) results in an effective zero bus voltage until the fault clears. Faults at least 1200 feet from the station result in a retained bus voltage that, depending on the clearing time, may cause either little or no disruption to adjacent customers during the event.
1 Tie Reactor - Bus A 0.9 0.8 Existing Feeder 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 1 2 3 4 5 6 7 8 Tie Reactor - Bus B Feeder Reactor

Distance from 23 kV Bus to the Fault (miles)

Figure 9-8 Bus Voltage (in Per Unit) During Line-to-Ground Faults as a Function of Distance From the 23-kV Bus to the Fault

To illustrate the effect of a tie reactor, two similar plots (shown in Figure 9-8) must be examined. The first curve (labeled Tie Reactor Bus B) shows the voltage on the bus for faults on the same side of the reactor as the fault. This curve also crosses zero retained voltage for faults at the fence line of the station. Faults further out tend to result in slightly worse bus voltage depression because of the reduced fault capability. Faults occurring on the side of the tie reactor opposite the bus under consideration (Tie Reactor Bus A in Figure 9-8) cause a much less severe sag. Faults as close to the station as possible (0 feet) result in a bus depression no greater than approximately 9-10

Custom Configurations

45%. This level of voltage depression is certainly low enough to cause equipment malfunction, but the distance from the station to the fault must be considered. The optimal placement of reactors as shown in Figure 9-8 is at each line position. The fault study considered line lengths out to approximately 7.5 miles. The curve labeled Feeder Reactor in Figure 9-8 shows that for the buses to be depressed to less than 80%, a line fault must occur within approximately 1 mile of the substation. In this study, a fault within approximately 2.5 miles without line reactors would cause a similar depression. Without consideration of the clearing time, and assuming 11 identical lines from the bus, the following results are found: Because there are 10 lines other than the faulted circuit, the total exposure is approximately 10 times 2.5, or 25 miles. When reactors are installed on each line, the result of the calculation drops to 10 milesan obvious improvement. Figure 9-9 further demonstrates improvements that are even higher than might be expected. represents the electrical distance from the bus for each segment of 35-kV class line. Although the line has more than 22,000 feet of conductor length, Figure 9-9 shows that a high percentage is located electrically nearer the bus. Lines that have a large number of spurs near the substation exhibit this characteristic. Because most of the faults on this line will be electrically near the bus and the greatest depth of the sag is for faults close to the station, the reactor offers an even greater improvement than initially expected.
% OF TOTAL CIRCUIT LENGTH

100 90 80 70 60 50 40 30 20 10 1000 2000 3000 4000 5000 6000 7000

ELECTRICAL FEET FROM BUS - END OF CIRCUIT 7700'

Figure 9-9 Percentage of Total Circuit Length Relative to Distance From Bus

Station Neutral Reactors and Customer Transformer Connections An inexpensive improvement for customer power quality is to use a neutral reactor on the substation transformer coupled with delta-wye transformers serving customers (see Figure 9-10). This can be used with any distribution system; it is often used on urban systems, especially those feeding secondary networks. For line-to-line-connected loads, the neutral reactor significantly improves voltages to end-use equipment, because of neutral shift. The reactor adds impedance, 9-11

Custom Configurations

which shifts the neutral point. This raises the voltage on the unfaulted phases but supports the line-to-line voltages. A very large reactor in a high-impedance grounded system would have almost no drop in the line-to-line voltages during a single line-to-ground fault.
Delta Grounded-Wye Substation With a Neutral Reactor

Delta Grounded-Wye Customer Connection

Figure 9-10 Neutral Reactor on the Substation Transformer Along With Delta-Wye Transformers

The lowest line-to-line voltage in per unit with a single line-to-ground fault is:
V= X 0 2 + X 0 X 1 + X 12 2X1 + X 0

Eq. 9-1

The neutral reactor of X ohms adds 3jX ohms to the zero-sequence impedance and raises the X0/X1 ratio. Table 9-1 shows that a modest neutral reactor significantly helps line-to-line loads during line-to-ground faults. Also, with delta-wye distribution transformers, the line-to-ground secondary voltages see equivalent sags as the line-to-line voltage on the primary. The neutral reactor provides benefit for only line-to-ground faults and no help for three-phase or line-to-line faults. A disadvantage of the neutral reactor is that it increases the voltage rise (the swell) on the unfaulted phases

9-12

Custom Configurations

Table 9-1 Line-to-Ground and Line-to-Line Voltages on the Low-Voltage Side of a Transformer With One Phase on the High-Voltage Side Sagged to Zero Voltages Primary Voltages Voltages Downstream of a Delta-Wye Transformer

No neutral reactor Line-ground Line-line 0.00 0.58 1.00 0.58 1.00 1.00 0.58 0.33 0.58 0.88 1.00 0.88

Neutral reactor giving X0/X1=3 Line-ground Line-line 0.00 0.72 1.25 0.72 1.25 1.00 0.72 0.60 0.72 0.92 1.00 0.92

Neutral reactor giving X0/X1=5 Line-ground Line-line 0.00 0.80 1.38 0.80 1.38 1.00 0.80 0.71 0.80 0.94 1.00 0.94

Conclusion The use of current-limiting reactors, although traditionally limited to short-circuit requirements, is also capable of affecting sag performance. Depending upon the placement of the impedance in the system, either a positive or negative consequence is possible. The effect has been shown when plotted on a distance-versus-fault plot. The ramifications of specifying higher-impedance transformers to limit fault current have the opposite effect of installing line reactors. Bus sags seen by all customers fed from the bus will experience greater depth sags. Neutral reactors are a cost-effective improvement under the right circumstances.

Advanced Customer-Service Configurations

Customer-Specific Solutions Utilities have several options for providing specific power quality solutions to customers; some of them discussed in this section include:

Subcycle medium-voltage transfer switches 1.5-cycle medium-voltage transfer switches Custom power devices Spot networks 9-13

Custom Configurations

These provide power quality benefits in addition to the traditional reliability improvements associated with traditional schemes. Depending on the sensitivity of the customers equipment, these designs might solve a customers power quality problem. Benefits A custom solution improves reliability and quality of power supplied to the customer. Negatives

These options may not completely solve problems for customers with extremely sensitive loads. Most options require two independent power sources, which can be costly and requires the utility to reserve capacity to serve the entire facility on both circuits.

Economics All are very expensive. Bottom Line Because costs are high, custom solutions must be used selectively. Also, all of the technologies must be tailored to the specific needs and sensitivities of the customers load. High-speed, mechanical transfer switches are a cost-effective solution if the load can withstand a 1.5-cycle interruption. A spot network provides excellent power quality with mature technologies. Other technologies may be appropriate depending on the loads characteristics. Reliability Versus Availability Utility performance assessment is based upon the statistical compilation of the number of customer interruptions and the duration of the interruptions. Tabulated on a seasonal or yearly basis, these indicators are useful for comparing the performance of distribution circuits. Maintenance efforts and capital improvements may then be more effectively directed. This results in overall cost savings and improved performance for both the utility and the customer base. Many industrial and commercial customers require power availability significantly greater than that provided by the standard radial distribution circuit. This need arises in some instances from public safety concerns such as the operation of fire pumps, from infrastructure requirements such as telephony installations, or from the need to maintain continuity of a process such as the operation of a high-tech manufacturing line. The specific needs are quite different from site to site. The customer may view service continuity as the most important criterion of electric service. Yet, utility reliability figures may be totally 9-14

Custom Configurations

inadequate to reflect actual service continuity. Often, momentary interruptions are not considered in reliability calculations if their duration is less than some minimum specified time. Successful instantaneous line reclosing following a self-clearing distribution fault is often not considered an event in the utility statistics. Design standards for ensuring reliability of industrial and commercial electrical service are principally based on ensuring that an alternative source of power is available should the primary feeder to the site be lost. For all practical purposes, the approaches for achieving this level of reliability are independent of a need to provide continuity of service. At least three means of providing power to critical loads exist:

The use of two or more feeders from the utility The use of on-site power in the form of either standby generators or on-line units operating in parallel with the utility The use of some form of uninterruptible power supply (UPS)a unit that has energy storage, traditionally in the form of batteries

These three approaches are entirely different and meet extremely different requirements. It is not uncommon to require the use of all three to meet the process requirements. Examples of Critical Installations Consider the telephone central station. For the telephone switching equipment, continuity of service is critical. Continuity is a term used to describe a power systems ability to supply continuous power with no interruptions. The telephone switching equipment itself uses direct current (dc) from banks of redundant batteries. Chargers, also redundantly configured, power these dc systems. The failure of any one charger or battery will not interrupt telephone service. Batteries have limited energy storage and are designed to provide energy during an interruption until a return of normal service or the transfer of service to an alternate meansa second utility source or a standby generator. Continuity of telephone service is maintained, regardless of the utility feeder conditionprovided that the generator starts successfully. The fire pump represents a different reliability requirement. A fire pump system will have backup systems in place and thus can survive a brief interruption in service. For this type of load, availability is critical. Availability is a term used to describe the ability of a power system to supply power when needed. If a manufacturing plant, for instance, uses a normally running fire pump, it probably has an engine-powered pump as a backup. In extreme cases, the fire pump system may also have an elevated tank that provides a minimum of water pressure even if all power is lost for a short duration. Continuity of service for fire protection is not critical. If the normal utility feeder trips, the engine will provide water pressure consistent with fire-fighting needs. Availability of service is paramount. This distinction between continuity of service and availability of electrical service is of utmost importance.

9-15

Custom Configurations

The requirements of the equipment for continuity and availability of service must be considered during the original design of the facility. The hospital will always provide for on-site emergency generation with perhaps a day of fuel supply. A second utility source would not eliminate this requirement. The hospital will also install a UPS on its data-processing center. These methods of enhancing reliability satisfy different needs. Most plants may assume that the process will be upset for any interruption of the electrical service. The importance of a second feeder or generator becomes that of providing a safe shutdown. Some loss of in-process material or equipment damage is assumed, and the safe recovery from the interruption is the overriding concern. The second source can provide continuity of service only for pre-arranged switching sequences. Manufacturing equipment has become less tolerant of even momentary interruptions. Directly driven processes that formerly used across-the-line motors are now controlled by adjustablespeed drives (ASDs). More power electronics have made the processes sensitive to what may have previously been brief, insignificant interruptions. Manufacturing is now more interconnected. Not only are the production lines now continuous rather than batch oriented, but the just-in-time concept also ensures that a glitch in the process between raw material and finished producteven if plants are across the world from each otherhas increased financial consequences. Customer Perception Data from customer surveys found in the IEEE Gold Book247 show that customers are at least aware that the duration of interruption is as important as the event itself. Chapter 1 of the standard, the Gold Book, tabulates the results of industrial customers as shown in Table 9-2.
Table 9-2 Critical Service Loss Duration for Industrial Plants Average Plant Upset for 1- to 10-Cycle Interruption 1.39 hours

Most Sensitive (25%) 10 cycles

Median (50%) 10 seconds

Least Sensitive (75%) 15 minutes

The commercial customer survey results are tabulated in a somewhat different format and are shown in Table 9-3.
Table 9-3 Critical Service Loss Duration for Commercial Buildings 1 Cycle 3% 2 Cycles 6% 8 Cycles 9% 1 Second 15% 5 Minutes 36% 30 Minutes 64% 1 Hour 74% 12 Hours 100%

IEEE Std. 493-1997, IEEE Recommended Practice for the Design of Reliable Industrial and Commercial Power Systems (Gold Book).

247

9-16

Custom Configurations

These surveys were taken before the relatively recent understanding of equipment sensitivity to sags. Although the survey respondents understood the effect of extremely short interruptions, it is unlikely that they based this understanding on actual monitoring or test results. Since the survey, a much greater understanding has been gained regarding the sensitivity of devices to momentary utility events. Surveys do show that customers are concerned with the length of the interruptions. The concerns for interruptions lasting as little as 10 cycles are probably in large part attributable to sags. Interruption and Sag Frequency
248 The EPRI Distribution Power Quality Study has been instrumental in the understanding of utility events and the development of many strategies to mitigate the end users problems. The distribution of sags and interruptions recorded during the study is shown in Figure 9-11.

Sag and Interruption Rate Magnitude Histogram

All Sites, One-Minute Aggregate Window 1.25 Sags and Interruptions per Site per 30 Days 100% 90% Cumulative Frequency 1.00 Sag and Interruption Rate Cumulative Frequency 0.75 80% 70% 60% 50% 0.50 40% 30% 0.25 20% 10% 0.00 10 to 15 15 to 20 20 to 25 25 to 30 30 to 35 35 to 40 40 to 45 45 to 50 50 to 55 55 to 60 60 to 65 65 to 70 70 to 75 75 to 80 80 to 85 85 to 90 5 to 10 0 to 5 0%

RMS Voltage Magnitude (%)

Figure 9-11 Sag and Interruption Rate From EPRI Distribution Power Quality Study

A voltage bar of 0 to 5% represents outages; the other bars represent sags. Sags are many times more likely than interruptions. The distribution of sag durations recorded at customer sites during a National Power Laboratory survey are shown in Figure 9-12249.
An Assessment of Distribution System Power Quality: Volumes 1-3, EPRI, Palo Alto, CA: 1996. TR-106294-V1, TR-106294-V2, and TR106294-V3.
248

9-17

Custom Configurations

Figure 9-12 Distribution of Sag Durations From a National Power Laboratory Study

The duration of the sags may be understood by realizing that some faults clear themselves without equipment operation, some are cleared by fuses (both expulsion and current-limiting), and others by reclosers or circuit breakers. The depth of the sag is related to the available fault current, nature of the fault, and distance from the station. Whole-Site Mitigation Power quality hardening would ideally be inherent to the design of all individual components. Equipment tolerance to sags is usually not specified and generally must be determined by aftermarket testing. The reliability of the entire system or process during utility sags is determined by the weakest link in the chain. Individual devices may be supported during sagsthe most common utility event. Alternatively, an entire line may be supported by the installation of a large-scale UPS. This device may provide conditioned power to an entire production line. Use of the UPS has the advantage of providing at least a small amount of energy storage and, when properly sized, will permit an on-site power generator to be started. The generator will continue to carry the entire process indefinitely. The disadvantages of the UPS include the additional maintenance and an inability to serve loads not physically close to each other. Where the ultimate in reliability is required, there is no alternative to providing an on-site energy system, such as a generator. Unless the generator is operated continuously in parallel with the utility,
Dorr, D.S., Hughes, M.B., Gruzs, T.M., Juewicz, R.E. and J.L. McClaine, Interpreting Recent Power Quality Surveys to Define the Electrical Environment, IEEE Transactions on Industry Applications, Vol. 33, No. 6, November 1997, pp. 14801487.
249

9-18

Custom Configurations

there is also no traditional alternative to the energy storage contained within the UPS to ensure continuity of service. The appeal of the whole-site power quality device is apparent for a number of reasons. Often the site is spread over a large area; no single point of installation, other than the utility tie, is practical. The plant process may contain an extremely large number of subsystems. These subsystems may be fed from different voltage levels and interact to the extent that correctly identifying where to place limited mitigation equipment is difficult. In addition, the performance of localized mitigation equipment cannot be demonstrated until after the fact. The final consideration is that process lines are frequently changed. Manufacturing methods are constantly improved, and the product of today may not be that of tomorrow. The production equipment itself is frequently changed. Standalone protection for computers will always be necessary, but whole-site protection is more practical because of advances in high-power switching devices. These devices may be intended to inject a series voltage that compensates for low voltage during sags or may provide a multi-megawatt off-line UPS. Both concepts have been employed at distribution voltage levels. Because of the limited energy storage, these devices are capable of supporting the normal site for a limited time measured in seconds. The voltage-injection devices are only designed to correct sags of a limited severitynot interruptions. The UPSs are able to support the entire load until a transfer to either a second utility source or an emergency generator occurs. Primary-Side Transfer Switches The requirement for higher electric reliability often plays a large role in site selection for industrial or commercial facilities. If two electrical sources are not already on the property, a second source is often installed. Two electrical sources are most often used in a normal/reserve configuration. A one-line diagram of a typical installation is shown in Figure 9-13.
Cir cu

irc

ui

t#

it # 2

Tie Breaker

Figure 9-13 One-Line Diagram of a Dual-Feed Transfer Scheme

The tie breaker normally operates in the closed position, and the entire plant is fed using Circuit #1. Should this circuit trip, its associated breaker opens, and the breaker to Circuit #2 9-19

Custom Configurations

closes. The utility generally specifies the trip settings and time delays. Site personnel usually accomplish the return to the normal source manually. The control schemes of the tie and the source breaker are electrically interlocked. When manually transferring load between circuits, the breaker control switch for the off-line service is operatedor the second service breaker is closed. Upon release of the control switch, the breaker that is not being closed opens. The two services are momentarily paralleled. The scheme may also use a low-side tie breaker depending upon the specific design. An uninterrupted transfer only occurs during manual control. Interruptions because of scheduled outages and retransfer are avoided, but during automatic operation, transfers cannot generally be faster than approximately 15 cycleseven if intentional time delays are eliminated. The possibility of losing both circuits still exists. This possibility requires the installation of onsite generation. The expense of generators and switchgear often limit the size of the unit. This limitation is overcome by segregating critical loads on a bus, usually at the utilization voltage (480 V, for example). The load may then be fed from either of the utility services or the emergency generator (see Figure 9-14).

irc ui t

Cir cu it

#1

#2

Tie Breaker

C
EMERG GEN

TRANSFER SWITCH Critical Load Bus

Figure 9-14 One-Line Diagram of a Dual-Feed Transfer Scheme With Backup Generation

Transfer switches in the configuration of Figure 9-14 are normally open-transition types. Open-transition means that the two sources are not paralleled during load transfer. The sensing of the loss of voltage is similar to that of the primary equipment. A time delay is provided to inhibit operation for momentary interruptions or sags. Once determined necessary, a start signal is sent to the generator. Fast-start generators are extremely capable of assuming full load after as little as 10 seconds. The transfer switch determines if this source is adequate and then switches the critical load bus to the generator. Upon return to normal, a time setting is also used, and often 9-20

Custom Configurations

retransfer is inhibited until personnel are available to reconfigure loads. Inhibiting retransfer prevents a second arbitrary interruption of load upon return to normal. A closed-transition transfer switch is only viable if the generator can be synchronized to the normal source and a means of verifying synchronization is necessary. Standard Transfer Switch Concerns
Load Faults

One extremely important consideration is response to faults on the load side of the transfer switch. The switch/breaker not only must be designed to withstand fault current for load-side faults but also must inhibit transfer during downstream faults and after upstream protective devices may have operated. Mechanical transfer switches may use breakers; the breakers provide overcurrent protection as a secondary function. Many transfer switches, however, are not intended for anything other than load operation. Regardless of the interrupting capability, the transfer switch should never transfer a fault between sources. This requires overcurrent sensing to determine where the voltage sag has originated.
Delayed Transfer

The time delay associated with the operation of both the normal/reserve breaker scheme and the generator transfer switch is intentional. Two issues are involved with the delaying of transfer. Breakers and transfer switches are not fast enough to transfer a load without the majority of connected equipment malfunctioning. Differences in load sensitivities may cause some equipment to stop while other equipment continues to operate for a short time. Rather than reenergizing various portions of the process at different or uncertain operability levels, it is preferred that all be de-energized and restarted in an orderly fashion. This prevents or limits potential damage.
Motor Transfer

A special class of problem is associated with motor loads. Motor residual voltage continues even after its power source is removed250. In effect, the motor becomes a generator for a short period. The voltage level, as well as the frequency (motor rpm), originally decays rather slowly when disconnected. As the speed slows, the relative position of the motor voltage vector falls increasingly behind the utility system phase. This vector will pass in phase with the utility vector as the angle increases to 360 and again at each multiple of 360. As the voltage decays, the length of the difference vector, when in phase, will increase with each successful rotation. As the voltage difference vector shows, at a 180 difference after disconnection, the voltage upon reenergizing will be practically twice the normal. On larger motors, closing at this voltage difference is extremely damaging to the motor. The National Electrical Manufacturers Association (NEMA) recommends that reclosing should be delayed for one time constant or

G.W. Bottrell, Hazards and Benefits of Utility Reclosing, in proceedings of Power Quality Conference, Oct. 1993.

250

9-21

Custom Configurations

more. The time constant is expressed in seconds and corresponds to the motor terminal voltage decaying to 36.8% of its initial value. The decay to this value depends on the motor load torque and the inertia of both the motor and the driven load. Figure 9-15 shows the relative position of the motor vector when compared to the fixed utility vector. Point #1 represents the utility. As the motor slows, the number identifying its phasor location increases.

Figure 9-15 Motor Vector in Relation to Fixed Utility Vector During an Interruption

A number of small motors may need to be considered as one large unit for some studies. The intentional time delay of the incoming breaker scheme and/or the transfer switch accommodates the motor residual decay. Of course, the generator start is usually inherently longer. (The value of one time constant is on the order of 0.2 to 10 seconds.) Retransfer from the generator to the utility is an extremely serious concern because the motor will be operating at speed and supplied from an unsynchronized source. Avoiding an out-of-synchronous transfer is one reason for inhibiting retransfer unless operator control is used. A second problem associated with motor residual voltage is the recognition of an interruption. Not every interruption to a customer results from a system fault. Occasionally, incorrect relay operation will open a feeder source. If an actual fault is present, the utility system may serve to load the motor terminals and therefore rapidly reduce the residual voltage. However, if the feeder 9-22

Custom Configurations

breaker or sectionalizing device opens, the monitor that controls transfer to the alternate service may have difficulty in rapidly detecting any problem. Figure 9-16 shows the effect at a power quality monitor when the customers line circuit breaker opened because of a blown fuse on a primary potential transformer.

60000 40000 20000

Voltage (Volts)

Phase A-B Voltage February 02, 1994 at 04:09:21

0 -20000 -40000 -60000 0 20 40 60 80 100

Time (mSeconds )

Figure 9-16 Motor Terminal Voltages at Start of a Power Outage

Figure 9-17 (a graph captured 6 seconds after the graph in Figure 9-16) shows both the sensing problem and the problem of re-energizing a motor with significant, residual terminal voltage.

9-23

Custom Configurations

note scale change

15000 10000 5000
Voltage (Volts)

February 02, 1994 at 04:09:27

6 Sec from tripping to 0 v

0 -5000 -10000 -15000 0 20

Time (mSeconds )

Freq 30 Hz
40 60 80 100

Figure 9-17 Motor Terminal Voltage at 6 Seconds After the Start of a Power Outage

Mechanical Wear
Delaying transfer creates a concern for mechanical wear. During widespread storms, frequent voltage sags may occur because of adjacent line faults. The mechanical life of the operating mechanism is definitely a concern.

Potential Changes to Existing Operation

Occasionally, the plant design will permit changes in the normal power supply that may benefit the customer. If process equipment is redundant, splitting the load between the two utility supplies may be useful. Systems are often designed where individual component redundancy is provided for failure, but not system redundancy. Are two separate cooling towers installed, each 100% redundant? Maybe. If both are run continuously and powered from separate utility feeds, the system might be tolerant of the loss of either utility feed. If three pumps are provided, but two are always required, the splitting of the utility source would not be beneficial. Two pumps would always be powered by one feed.

Incorrect splitting of plant loads may result in a significant decrease in plant availability. Unless the loads are 100% redundant, placing portions of either system on the second service will make them vulnerable to an event on either supply. Rather than performance being enhanced, it will be worsened. Ultra-Fast Primary-Side Transfer Switches The same high-power, solid-state switching devices that make the UPS a reality are also being 251 252,253,254,255 for more offered as an alternative solution (see EPRIs Static Switch Primer and
Static Transfer Switch Primer, EPRI, Palo Alto, CA:1998. TR-111697. D. Dull and L. Sabados, Application of Medium-Voltage, 1.5-Cycle Switching for Critical Loads, in proceedings of Power Quality Conference, Chicago, 1999. 253 Descriptive Bulletin 650-32, Pure Wave Source Transfer System, S&C Electric Company.
252 251

9-24

Custom Configurations

information). If the customer site has two utility sources, a subcycle transfer switch becomes practical. If the sensing speed plus the transfer speed is under a cycle, no concern regarding motor residual voltage exists. The typical static switch operates in less than cycle, and cycle is claimed. This speed is fast enough that it is highly unlikely any equipment will be affected. It is important that the alternate source is independent so that the voltage sag experienced by the normal source does not also appear on the circuit acting as the reserve feeder. Figure 9-18 shows a typical static-switch configuration, and Table 9-4 lists manufacturers.

Figure 9-18 Typical Static Switch Configuration Table 9-4 Ultra-Fast Primary-Side Transfer Switches256 Manufacturer InverPower Mitsubishi Name STS-D LightSpeed SSTS Silicon Power Corporation (SPCO) S&C Electric Corporation PowerDigm MVSTS PureWave STS Voltage Rating 15 kV44 kV 4 kV35 kV 5 kV38 kV 4 kV35 kV Device Rating 200600 A 200 1200 A 400 1400 A 300 1200 A Transfer Time 210 ms 4 ms Web-site http://www.inverpower.com http://www.meppi.com

4 ms8 ms 2 ms4 ms

http://www.siliconpower.com

http://www.sandc.com

Figure 9-19 shows the importance of circuit independence. If both service feeders originate from the same bus, the sag seen on Feeder B will be the same as that at the bus. Even with a subcycle

R.E. Tyner, A High-Speed Transfer System Utilizing Medium-Voltage Power Circuit Breakers, in proceedings of Power Quality Conference, Oct. 2000. 255 Guidebook on Custom Power Devices, EPRI, Palo Alto, CA. 2000. 1000340. 256 Guidebook on Custom Power Devices, EPRI, Palo Alto, CA. 2000. 1000340.

254

9-25

Custom Configurations

transfer, the alternate source may not be adequate to prevent equipment malfunction. The following list describes preferred situations for fast-switch installations:

Preferred and alternate feeders fed from different substations: This configuration allows the utmost protection. For example, if the preferred feeder substation trips, the alternate feeder will not lose power. One possible source of failure for this configuration is if the transmission system feeding both substations has a disturbance. This situation would obviously result in both feeders experiencing the disturbance. Preferred and alternate feeders fed from different substation transformers: If both feeders are fed from the same substation, using different substation transformers offers increased isolation for each feeder. If both feeders are fed from the same transformer, a fault on either feeder could cause the transformer circuit breaker to trip, resulting in a loss of both feeders. Using two transformers allows the transformer feeding the fault to be taken off line without affecting the unfaulted feeder. Different ROW for preferred and alternate feeders: Using different ROW (right-of-way) increases the independence of system disturbances on the feeders. Assuming that the same ROW was being used, and a fault occurs on the line (downed pole, fallen tree, and so on), the probability of both feeders experiencing the fault increases considerably.

Fault studies are essential in determining the independence of each feeder during system disturbances. These studies are usually available from the utility. Fault data from previous years can also offer insight into the dependence of parallel feeders.

2.5 s

2.5 s

Distribution Substation

Substation Monitor

Recloser

Feeder Monitor

CB
Feeder A Circuit Breakers
Substation Monitor

x
Fault

x
Feeder B

CB

2.5 s

2.5 s

Figure 9-19 Voltage on Two Feeders Served From a Common Bus With a Fault on One Feeder

Ultra-fast transfersunlike delayed transfersare intended to maintain all loads as before the transfer. Ultra-fast transfers are unlike delayed transfers, where a large portion of the site load 9-26

Custom Configurations

may drop off-line. An uninterrupted transfer will transfer the entire load to the second distribution circuit. If the plant load is significant, distribution voltage drop may be a concern, because the position of voltage support devices such as regulators and capacitor switches requires a minimum time to readjust to the increased loading. FastBut Not SubcycleTransfer Switches Various equipment sensitivities to voltage sag depth and duration have been established for generic classes, as well as a multitude of specific components. The fact that not all devices are equally intolerant of sags has resulted in a compromise device for whole-site mitigation. These devices are called high-speed transfer switches. They use vacuum switch technology in the power circuitry and subcycle event detection similar to the -cycle static-transfer switch technology. The operational speed of this device is approximately 1.5 cycles. Vacuum switches must wait for a zero current crossing to interrupt the load before closing the alternate-side device. Transfer to a synchronized source requires no intentional delay. The value of either the subcycle or high-speed switch is dubious if the second source is not within a limited range of phase difference versus the normal supply. The products based on vacuum technology may fill a void where the expense of the static switch is excessive or where it is desired to use more conventional technology. Table 9-5 gives basic information about two high-speed mechanical transfer switches.
Table 9-5 Manufacturers of High-Speed, Medium-Voltage, Mechanical Transfer Switches Voltage Rating (kV) Joslyn Hi-Voltage Corporation Powell Industries, Inc. FasTran25 Cycle+ 427 5.527 Device Rating (A) 600 600 25 22 http://www.joslynhv.com http://www.powellesco.com Transfer Time (ms)

Manufacturer

Product

Web Site

The least sensitive models of ac relays tested would drop out at slightly more than 30 cycles. Models tested subsequently have revealed even greater sensitivities. Plug-in relays have been found extremely intolerant of sags. Some devices while operating indefinitely at low voltages near 60% of nominal voltage will drop out as a result of a sag to 50% of nominal voltage for even a -cycle sag. Process systems are a collection of subcomponents. The malfunction of any of these is usually sufficient to cause the entire operation to shut down or sustain damage. The time for operation of the high-speed vacuum devices overlaps the interval where many devices drop out. This requires a serious evaluation of the weakest link in the process and providing local mitigation where indicated. This evaluation usually involves on-site testing and re-testing because the most sensitive devices are sequentially supported. The operation of the subcycle transfer switch is so fast that there is no concern.

9-27

Custom Configurations

Another particularly sensitive device is the high-intensity discharge bulb. The specific sensitivities found were not only related to type but also to age. Lighting response to high-speed transfer switch operation is a definite area for concern. An obvious tradeoff with the use of a high-speed versus a subcycle transfer switch is that all loads will see the interruption time during switching. During single-phase utility faults, some energy conversion is still possible in three-phase equipment; this may permit continued operation during single line-to-ground faults. Some single-phase loads may not even be subjected to any sag, depending on the transformer connections. In these cases, the three-phase interruption, no matter how brief, may actually be more disruptive than riding through the sag. The benefits of transferring during sags and introducing a three-phase interruption instead of letting equipment ride through require a thorough analysis. The effect on utilization equipment for primary faults is shown in Figure 9-20 and Figure 9-21. Figure 9-21 shows a single line-to-ground fault on a feeder to a delta-to-wye-connected transformer. The per-unit voltages are tabulated. The case for a wye-wye connection is somewhat easier to visualize. The remaining voltage during a sag extends to the 480-208/120-volt transformation, which is usually delta-wye.
A PHASE PRIMARY B PHASE PRIMARY
SECONDARY 1 pu - -N

C PHASE PRIMARY

NORMAL PHASORS

Figure 9-20 Normal Voltage Phasors for a Three-Phase Service

9-28

Custom Configurations
A PHASE PRIMARY B PHASE PRIMARY

C - SLG FAULT TRANSFORMER C BUSHING TIED TO GROUND

A B

PHASORS DURING SLG FAULT (Phase C)

Figure 9-21 Voltage Phasors for a Single-Line-to-Ground Fault on Phase C

The immediate inadequacy of the traditional normal/reserve electrical supply scheme, regardless of breaker timing, is the means of sensing a problem. Even if instantaneous transfer were possible, undervoltage relaying would be inadequate. The loss of source voltage sensing is often electromechanical in design and more suitable for root-mean-square (RMS) rather than point-onwave sensing. To avoid unnecessary breaker operation, an intentional time delay is added to permit the line to return to normal. This philosophy is adequate for ensuring availability of service but is inconsistent with requirements to provide continuity of process. The successful transfer of load, without upset, first requires that voltage comparisons be made on a point-bypoint basis. It is not sufficient to detect a loss of a phase entirely; transfer must be considered if the waveform departs from the desired ideal wave shape. This requirement is not surprisingly unlike that required to trigger power quality monitors. Because connected devices can malfunction for PQ events lasting less than a cycle, sensing voltage quality should be made on an order of magnitude faster than any expected response. This sensing ability applies for both the solid-state subcycle switch and the electromechanical high-speed switch. The equipment sensitivity will determine whether the slower switch is adequate. Any transfer equipment by its nature creates a new common point of failure. This may be because of the location of a single device in the same cabinet or may involve only the cabling between the two units. Any single failure that could affect both sources cannot be permitted. Similarly, provisions for maintenance must be provided. The devices must be designed for bypassing, manually as a minimum and automatically if required. Each device must be repairable without affecting the load. Even a catastrophic failure should not affect the availability of a power source, ideally not even affecting the continuity of service. High-speed, mechanical transfer switches are approximately $150k compared to $500k for a static switch. Additional costs include building the second feeder to the customer location and providing reserved capacity on the backup feeder. While high-speed mechanical switches require no more maintenance than a typical automatic-throwover scheme, a static switch may require supplemental cooling.

9-29

Custom Configurations

Secondary-Side Transfer Switches Secondary-side transfer designs have many of the same advantages and disadvantages of primary-side designs. Secondary transfer designs require two independent sources. Subcycle secondary transfer switches are available with performance characteristics similar to subcycle primary transfer switches. The advantages and disadvantages of secondary designs are:

Transformer failure: A secondary switch arrangement can supply load if a transformer or associated equipment fails. Space: Because of the redundant transformer, the secondary-side arrangement takes up more space. Cost: A secondary-side arrangement typically costs more than a primary-side arrangement.

Custom Power Equipment Custom power is the employment of power electronic or static controllers in distribution systems rated at 1 kV through 38 kV for the purpose of supplying a level of reliability and/or power quality that is needed by electric power customers who are sensitive to power variations (see EPRIs Guidebook on Custom Power Devices 257 for more information). Custom power devices, or controllers, include static switches, inverters, converters, injection transformers, mastercontrol modules, and/or energy-storage modules that have the ability to perform currentinterruption and voltage-regulation functions in a distribution system. The main types of custom power equipment are:

Static var compensator (SVC): Shunt device that dynamically injects or absorbs reactive power into the system, usually with a thyristor-switched reactor or capacitors; best for flicker control. Static shunt compensation: A shunt-connected, solid-state switching power converter that exchanges reactive current with the distribution system; has the same features as an SVC but can also filter harmonics. Static series compensator: A series device that mitigates the effect of voltage sags and interruptions by injecting a signal that helps hold up the voltage. Static voltage regulator: A step-voltage regulator but with taps quickly controlled by thyristor switches. Backup energy supply devices: These are an uninterruptible power supply applied at medium voltagea battery or other energy-storage technology through a power electronic package provides energy to loads during voltage sags or momentary interruptions.

Custom power devices are expensive and must be matched with the characteristics and sensitivities of customer loads. Table 9-6 shows typical costs and features. These are typically used for only very sensitive customers with demands for high-quality power.

257

Guidebook on Custom Power Devices, EPRI, Palo Alto, CA. 2000. 1000340.

9-30

Custom Configurations

Table 9-6 Custom Power Device Application Matrix258

Source Transfer Switch Static Voltage Regulator Backup Energy Supply Device

Power Quality Problem

Static VAR Compensation 1

Static Shunt Compensation 3

Static Series Compensator 5

Voltage sag Voltage swell Momentary interruption CapacitorSwitching Transient Voltage regulation Flicker Harmonics Reactive Power Compensation Cost range (U.S. $)
1

50200/kvar

120175/kvar

500 1000/am p

150250/kVA

80 125/kVA

750 1500/kVA

Generally corrects up to 100% of nominal for a sag to 65 to 70% of nominal. Corrects up to 90% of nominal for a sag to 55 to 60%. 2 One model corrects 5th and 7th. Other two models have de-tuning filters provided for harmonics determined to be present. 3 Generally corrects up to between 90% and 100% for a sag to 20% of nominal. 4 One model corrects 25th harmonic, the other corrects to the 17th. 5 One device corrects for voltage sags to 0%. Others typically correct for voltage sags down to 50%.

Spot Networks Especially in urban areas, utilities often use spot networks to feed major loads, such as a highrise building. Normally, spot networks are used in urban settings, but they can be used anywhere. The spot network is generally fed by two to five primary feeders that are connected to the lowvoltage network via network transformers (see Figure 9-22). The secondary is paralleled; all feeds are tied together. By installing network protectors between the secondary terminals of the network transformers and the grid network, faults on any of the parallel paths feeding the grid are isolated without causing an outage to the load served from the network. Similarly, a parallel path can be taken out of service for maintenance or modification, and the loads served from the cable grid do not experience an interruption. Network protectors in normal applications have a sensitive reverse-power-trip setting that disconnects the unit from the faulted source. Spot networks supply excellent reliability. When one or more feeders is lost, the remaining feeders can carry the load as long as the remaining transformers are sized to carry the load. We can
258

Guidebook on Custom Power Devices, EPRI, Palo Alto, CA. 2000. 1000340.

9-31

Custom Configurations

design the spot network to withstand a loss of one circuit (single contingency or N-1) or a loss of two circuits (double contingency or N-2). A spot network is a very reliable supply system in terms of short- and long-duration interruptions. Mean times to failure of 10 to 30 years are normal. They can also provide some but not complete protection against voltage sags.

Network transformer Network protector

or

208Y/120 V or 480Y/277 V spot network

Figure 9-22 A Spot Secondary Network

Spot networks have advantages over transfer switches:

Redundancy: Whereas transfer switch designs have two supplies, spot networks can easily provide more redundancy, more feeder circuits. This is very difficult to do with switch schemes. Equipment complexity: Spot networks do not require electronic equipment or other immature technologies. The network transformer and network protector are mature, proven, and reliable. No interruption during transfer: Spot networks do not transfer, so they do not ever interrupt the load (although there will be a voltage sag until the station breaker clears the fault). Cost: A spot network is less expensive than a primary- or secondary-side static switch. Of course, spot networks are still expensive and require one or more additional primary supplies.

The main performance disadvantage of spot networks is that they do not provide significant protection against voltage sags. If a fault occurs on a primary circuit, the substation bus voltage will sag, and because all feeders are from the same substation, the spot network and the customer see low voltage. Once the station breaker opens, the voltage will partially recover and then fully recover once the network protectors open on the faulted circuit. A spot networks main defense against voltage sags is fast operation of the substation breaker. Any measures to improve the 9-32

Custom Configurations

operating speed of the relaying and breakers will improve performancetwo or three cycle breakers with an instantaneous relay element. With the right transformer connections, spot networks can be much less sensitive to voltage sags. If the substation transformer has a modest grounding reactor, and the network transformer has a delta-to-grounded-wye connection, then the spot network will not see voltage sags from line-toground faults on the primary circuits. See Section 7 for more information. Another option to improve voltage-sag performance is to add reactors as shown in Figure 9-23. The reactors reduce the impacts of faults on primary feeders (those between the reactors). The reactors add more impedancethis provides isolation, so the voltage sag is much less severe at the spot network. The addition of the reactors will introduce some potential problems that would need to be addressed such as effects on voltage drop and possible changes in coordination of protective devices. Reactors would also limit the fault current that can be a problem at some urban substations. Note that the reactors do not help with sags caused by transmission-level faults.
Traditional Spot Network PQ-Enhanced Spot Network

Network transformer Network protector

or

or

208Y/120 V or 480Y/277 V spot network

208Y/120 V or 480Y/277 V spot network

Figure 9-23 Comparison of a Traditional Spot Network to a Configuration With Added Reactors to Minimize the Voltage Sags Caused by Primary Feeder Faults

Summary Most of these customer-specific options are targeted solutions for especially critical and sensitive customers. No one solution provides complete power quality support. Table 9-7 compares the performance and costs of some of the most prominent utility-side solution options. Also remember that utility-side options can be used in conjunction with customer-side options, either at the facility level or at the equipment level.

9-33

Custom Configurations Table 9-7 Comparison of Customer-Specific Configurations Performance LongDuration Interruptions Slow switch Subcycle switch 1.5-cycle switch Static series compensator Medium-voltage UPS Spot network Good Good Good None None Very good ShortDuration Interruptions None Good Good Good Very good Very good Voltage Sags None Good Good Good Very good Fair

Maturity Very mature New New New New Very mature

Cost Moderate Very high High High Very high High

9-34

10
SPECIFIC POWER QUALITY ISSUES
The majority of this report is focused on voltage sags and interruptions but there are certainly many other power quality issues that impact customers. This chapter aims to provide and introductory looks at some of these other power quality issues as well as some methods for solving them.

Voltage Regulation
ANSI standard C84.1 specifies the preferred voltage levels for electric power systems. Most utilities have adopted the C84.1 standard as a minimum requirement, and some have more stringent requirements imposed by their respective public utilities commission. The C84.1 standard has two ranges: Range A and Range B service voltages259. These are defined as follows:

Range A: Proper range considered to be within limits Range B: Infrequent excursionThe occurrence of service voltages outside of these limits should be infrequent and When they occur, corrective measures shall be undertaken within a reasonable time to improve voltages to meet Range A requirements.

The standard gives ranges at two locations, the service voltage (voltage at the point where the electrical system of the supplier and the electrical system of the user are connected) and utilization voltage (voltage at utilization equipment). The C84 ranges are shown in Figure 10-1.

259

ANSI C84.1-1989, American National Standard for Electric Power Systems and Equipment - Voltage Ratings (60Hz).

10-1

Specific Power Quality Issues

Voltage (120-V Base)

104 108 112 116 120 124 128

Utilization Voltage Service Voltage: 120- to 600-V Systems Service Voltage: Systems 600 V or Greater

Range A

Utilization Voltage Service Voltage: 120- to 600-V Systems Service Voltage: Systems 600 V or Greater Nominal System Voltage

Range B

The shaded portions on the lower end of the utilization voltages do not apply for circuits supplying lighting loads. The shaded portion at the top Range A utilization voltage does not apply to 120- to 600-V systems. Figure 10-1 ANSI C84.1 Voltage Ranges

Sustained overvoltages or undervoltages can cause the following end-use impacts:

Tripping of sensitive loads: For example, an uninterruptible power supply (UPS) may revert to battery storage during high or low voltage. This may drain the UPS batteries and cause an outage to critical equipment. Improper or less-efficient equipment operation: For example, lights may give incorrect illumination or a machine may run fast or slow.

In addition, undervoltages can cause overheating of induction motors. For lower voltage, the induction motor will draw higher current. Operating at 90% of nominal, the full load current will be 50 to 10% higher, and the temperature rise will be 10 to 15% higher. Reduced motor starting torque also occurs. Also, overvoltages can cause:

Equipment damage or failure: Equipment can suffer insulation damage. Light bulbs wear out much faster at higher voltages. Higher no-load losses in transformers: Magnetizing currents are higher at higher voltages.

Solutions to Voltage-Regulation Problems The main solutions to feeder-wide voltage regulation problems are: 10-2

Specific Power Quality Issues

Check existing equipment: Make sure that all existing regulation equipmentsubstation LTC controls, switched capacitors, line regulatorsis performing properly. Also check to see if settings can be modified to improve performance: Raise voltage setpoints or use more aggressive line-drop compensation. Shift the load: Reduce the load by moving some loads to another circuit. Add switched capacitors: Switched capacitors help improve voltage profiles. As a side benefit, capacitors will normally reduce line losses. Using a capacitor controller with voltage control provides the best control of voltage. Add voltage regulators: Line regulators will improve voltage regulation.

More extreme and more expensive solutions are upgrading to a higher voltage, reconductoring a line, or undergrounding a circuit. The power-flow program is the main tool to use to evaluate changes in loading or adding equipment. Spot measurements, substation loading information, and other meter-type information can help fine-tune models. Before implementing feeder-wide solutions, make sure that the problem is not isolated to one customer. Check for common problems like an undersized transformer or an open neutral.

Voltage Unbalance
In three-phase power systems, the generated voltages are intended to be sinusoidal and equal in magnitude, with the individual phases 120 apart. However, the resulting power system voltages at the distribution end and at the point of utilization can become unbalanced. The nature of the unbalance includes unequal voltage magnitudes between phases at the fundamental system frequency (phase voltages unequally above or below system nominal voltage), differences in phase relationship between the phases, and unequal levels of harmonic distortion between the phases. Most distribution systems use four-wire configurations with single-phase distribution transformers connected line-to-neutral to serve single-phase loads, such as residences and street lighting. Variations in single-phase loading of the phases cause unequal currents and therefore unequal voltage drop in the three phases. This unbalances the receiving end-phase voltages. Additional causes of power system voltage unbalance can be asymmetrical transformer winding impedances, open wye and open delta transformer banks, asymmetrical transmission impedances possibly caused by incomplete transposition of transmission lines, malfunction of voltage regulators resulting in a stuck regulator on a particular phase, and blown fuses on three-phase capacitor banks. Example problem areas can be rural electric power systems with long distribution lines, as well as large urban power systems with heavy single-phase demands, such as street lighting and residential loads fed by single-phase laterals. Single-phase traction in electric transit and railroad systems can also cause considerable unbalance on the utility three-phase system unless proper design steps are taken. Industrial and commercial facilities may have well-balanced incoming supply voltages, but unbalance can develop within the building from its own single-phase power requirements if the 10-3

Specific Power Quality Issues

loads are not uniformly spread among the three phases. Within a user facility, unbalanced voltages can also be caused by unbalanced and overloaded equipment and high-impedance connections such as bad or loose contacts. An example of unbalanced equipment is motor impedance unbalance, which has been seen to increase with time, possibly because of the unbalanced heating of the stator. Motor-caused unbalance can be due to a problem in manufacturing such as unequal number of turns in the windings, a misaligned rotor, or an asymmetric stator. Motor-winding unbalance can also occur in the repair process where failed windings are quickly and inexpensively repaired by isolating the failed turn, thus reducing the impedance of the repaired phase. There are several ways to measure and calculate unbalance. The method most commonly used in the U.S. is the National Electrical Manufacturers Association (NEMA) Standards Publication No. MG 1-1993, which defines the voltage unbalance in percent as:
% Unbalance =

Maximum Deviation from Average 100 Average of Three Phase - to - Phase Voltages

Eq. 10-1

Note that line-to-line voltages are used in this NEMA standard as opposed to line-to-neutral voltages. When line-to-neutral voltages are used, phase angle differences are not reflected in the calculation. Therefore line-to-neutral voltages are seldom used to calculate voltage unbalance. Another useful measure, defined in European standards, indicates the degree of unbalance by a voltage unbalance factor (VUF), which is the ratio of the negative-sequence voltage to the positive-sequence voltage, represented as:
% VUF =

V2 100 V1

Eq. 10-2

V1 and V2 are the positive and negative sequence voltages, respectively, and can be obtained using symmetrical components analysis. Figure 10-2 presents the approximate percent voltage unbalance on the U.S. distribution system obtained from field surveys and reported in the American National Standards Institute (ANSI) Standard C84.11995 for Electric Power Systems and Equipment Voltage Ratings (60 Hertz). Many utilities do not keep track of their voltage unbalance in the interest of time and task prioritization. This is because the balance is generally quite good, and the adverse effects are not immediately apparent. Unbalance is thus typically only addressed when there is a complaint. Power quality surveys have tended to focus on disturbances involving voltage events such as sags, swells, overvoltages, undervoltages, impulses, transients, surges, and interruptions (outages).

10-4

Specific Power Quality Issues

100

98

% of U.S. Distribution Systems

80 66 60

40

20 2 0 <1% <3% >3%

% Voltage Unbalance

Figure 10-2 Approximate Percent Voltage Unbalance on the U.S. Distribution System

The main impact of voltage unbalance is on induction motors, which represents one of the single largest end-use load segments. The negative-sequence voltage resulting from voltage unbalance produces an air-gap flux rotating against the rotation of the rotor, thus generating an unwanted negative (reversing) torque. The result is a reduction in the net torque and speed and the possibility of speed and torque pulsations and increased motor noise. In addition, the negativesequence component in the unbalanced voltages generates large negative-sequence currents due to the low negative-sequence impedance. This can increase the machine losses and raise operating temperatures. At normal operating speeds, unbalanced voltages cause the line currents to be unbalanced in the order of 6 to 10 times the voltage unbalance. Overall, the net effect of the voltage unbalance is reduced efficiency and decreased life of the motor. Voltage unbalance as low as 2% can cause customer equipment to trip on current unbalance. This is especially the case for many chiller compressor motors, which include negative-sequence protection based on the current unbalance. This protection is typically set at 10% current unbalance, which may result from a voltage unbalance in the range of 1 to 2%. In addition to motors, voltage unbalance causes three-phase power electronic loads such as adjustable-speed drives (ASDs) to draw unequal current in each phase. This effect is more pronounced in smaller drives (<15 hp), where the percent current unbalance may be 20 times the percent voltage unbalance. High-phase currents can also result in over-current trips and excessive dc bus ripple. High ripple can reduce the life of the dc bus capacitor of the ASD.

10-5

Specific Power Quality Issues

Solutions to Voltage Unbalance Diagnosing voltage unbalance is easya voltmeter across each phase quickly tells the field engineer how severe the problem is. Unbalance problems are not usually intermittent; unbalance may fluctuate with load cycles, but it is fairly regular. Finding the source of the unbalance problems can be harder but still straightforward. Unbalanced loading between phases is the usual culpritmeasure the current on each phase to identify unmatched currents. The main solution to voltage unbalance is to balance up loads. The main problem areas are:

Large single-phase taps: Load is most easily shifted around at large single-phase taps. Very heavily loaded single-phase taps may be candidates for upgrades to two-phase or three-phase lines. Middle phase: Crews often avoid the middle phase out of convenience, especially on tight structures. Make sure to load the middle phase as much as the outer phases. Capacitor fuses: Check for fuses blown on capacitor banks (Figure 10-3). Capacitor fuses can blow from failure of the capacitor or other equipment, and many utilities have also had many nuisance fuse operations when the equipment is still good (the causes are unknown). Regardless of the cause, unbalanced capacitor banks can introduce significant voltage unbalance. Voltage regulators: Regulators that do not work or are set wrong can also cause unbalance. The load can be perfectly balanced between phases, but if one regulator is stuck, the voltages may become unbalanced. If single-phase regulators are set differently (either the voltage setpoint or the line-drop compensation settings), the set of regulators can unbalance the voltages. This becomes obvious after checking voltages upstream and downstream of a regulator.

Figure 10-3 Fuse Blown on a Capacitor BankA Common Cause of Voltage Unbalance

10-6

Specific Power Quality Issues

Voltage Flicker
Light flicker is due to rapidly changing loads and generation that cause fluctuation in the secondary voltage. Even a small change in voltage can cause noticeable lamp flicker. The degree of irritation depends on the frequency as well as the magnitude of the fluctuations, as indicated in the classic flicker curves shown in Figure 10-4, which depict the threshold of human annoyance. Most people will notice 1% and smaller voltage changes that occur in the range of 1 to 33 changes per second (60 to 2000 changes per minute). The greatest sensitivity occurs around 900 changes per minute, or 15 changes per second, which is equivalent to a frequency of 7 to 8 Hz. Both incandescent lamps and fluorescent lamps are susceptible to voltage fluctuations. The amount of flicker also depends on the type of lamp. Voltage fluctuations can also cause televisions and computer monitors to waver.

Figure 10-4 GE Flicker Curve

Annoying lamp flicker can occur when rapid changes in load current cause the power system voltage to fluctuate. Both incandescent and fluorescent lamps can flicker during voltage fluctuations. The standards for measuring and limiting lamp flicker are based on the 60-W incandescent lamp. Flicker is a difficult problem to quantify and to solve. The untimely combination of factors that lead to a flicker problem are: 1) some deviation in voltage supplying lighting circuits and 2) a person being present to view the possible change in light intensity due to the voltage deviation. The human factor significantly complicates the issue, and for this reason flicker has historically been deemed a problem of perception. The voltage deviations involved are often much less than the thresholds of susceptibility for even the most sensitive electronic equipment. System operating problems are only experienced in rare cases. To observers in offices, on the other hand, 10-7

Specific Power Quality Issues

voltage deviations on the order of 1 to 2% could produce extremely annoying changes in light output, especially if the range of repetitive deviations is 5 to 15 Hz. Due to the clear coupling of voltage fluctuations and lamp output changes, the term flicker often means different things to different people, with the interpretation primarily governed by point of view. In each case, the deviation may or may not be strictly periodic and is usually expressed as a change relative to the steady-state level. For voltage variations, the change is usually expressed as V/V. A similar expression for light intensity variations also exists. In contrast to the behavior of incandescent filament-type lamps, fluorescent gaseous discharge lamps have very little thermal inertia and respond even faster. While the time constant for a 60W, 120-V incandescent lamp is about 28 ms and a 230-V lamp of the same wattage about 19 ms, a typical fluorescent lamp has a time constant of less than 5 ms. Also, fluorescent lamps are usually not amplifying the voltage changes in their corresponding change in light output. This is why incandescent lamps are historically the most sensitive and are used to set standards for allowable voltage fluctuations. Flicker prediction and measurement can be challenging. The IEEE flicker curve, which was first developed by General Electric in the 1920s, provides a good calibration of the threshold of where voltage fluctuations at a specific frequency will cause visible lamp flicker. However, there is no IEEE standard on how to measure the flicker level in a complex varying voltage. Because the IEC does have an effective measurement standard, the IEEE working group on flicker recently voted to adopt the IEC standards. The number of this standard is IEC 61000-4-15, and more 260 information about this adoption is available in an IEEE committee paper . Motor Starting Motor starts are a special case of voltage flicker. Most motors start only a few times per day, which is possibly off the flicker curve. While motors may start infrequently, when they do start, the voltage change is sharp, can be deep, and can easily be visible. Motors normally draw five to six times the full-load current during starting. Normally, the voltage drops suddenly and then gradually recovers over several seconds as the motor comes up to speed. Utilities normally have a criterion for motor starts based on the change of voltage and how often the customer starts the motor. Willis261 reported that many utilities use a criterion of 3% during motor starting, but some utilities vary on how they define the percentage (either being relative to the nominal voltage, the minimum voltage, or the voltage at the time of the motor start). Utilities also often limit the size of motors or the starting current allowed by end users depending on the voltage.

Halpin, S.M., Bergeron, R., Blooming, T., Burch, R.F., Conrad, L.E. and T. Key, Voltage and Lamp Flicker Issues: Should the IEEE Adopt the IEC Approach? IEEE P1453 working group paper. http://www.manta.ieee.org/groups/1453/drpaper.html. 261 H.L. Willis, Power Distribution Planning Reference Book, Marcel Dekker, Inc., 1997.

260

10-8

Specific Power Quality Issues

Solutions to Voltage Flicker Diagnosing flicker is easya visual check of lighting is often enough. Often, the fluctuations are regular, but sometimes, flicker comes and goesit mainly depends on how often the fluctuating load runs. Tracing the source of the fluctuations can be tricky. Following the path of current fluctuations will lead to the source of the problem. And, voltage flicker gets worse closer to the fluctuating load. While this sounds easy, it can be difficult, especially if the flicker source is intermittent. Utility-side flicker solutions can be difficult and expensive. Load-side solutions are often more cost-effective. Some of the most common utility-side solutions are:

Connecting the fluctuating load to another circuit Upgrading the circuit voltage Running a dedicated circuit to the fluctuating load Using a static var compensator (SVC) Adding a series capacitor

All of these options usually require significant expense. Static var compensators (SVCs) dynamically change their reactance, injecting or absorbing reactive power to regulate the voltage in real time. They are often used at large arc furnaces to control voltage flicker, but distribution-scale versions are available. Several styles of var compensators are available:

Switched inductor: A thyristor controls how much reactive power is drawn from the system depending on where on the voltage waveform the thyristor turns on. Full vars are drawn if the thyristor turns on when the voltage is at its peak (a full half cycle of current is drawn). Fewer vars are drawn if the controller waits to turn on the thyristor. Once fired, the thyristor stays on until the current goes through zero and the thyristor shuts off. A drawback to this configuration is that the current blips are not sinusoidal and create harmonics. The inductor may be used in parallel with a shunt capacitor to control vars negatively and positively. Switched capacitors: In this configuration, several shunt capacitors each have thyristors to control whether the capacitor is connected. The number of capacitor units switched in determines the amount of vars injected. Having more individual blocks of capacitors helps keep the control smooth. The switched capacitors may be paralleled with a fixed inductor to allow the unit to control vars in both directions. STATCOM: The most advanced reactive power controller, this custom power device has a power electronics interface (normally a pulse-width-modulated inverter using IGBTs or other fast-switching electronics) that injects or absorbs reactive power that is temporarily stored in capacitors on a dc bus. The STATCOM can smoothly vary the reactive power with little harmonics. The STATCOM may be extended to inject or absorb real power with auxiliary energy storage.

10-9

Specific Power Quality Issues

None of these devices can instantly counter changes from fluctuating loads; some delay is always present. The delay is normally on the order of one to two cycles, depending on the compensator and controller technology. To fully offset the drop, the kvar capability must approximately equal the motor starting kVA. Distribution static var compensators are about $50 to $250/kvar. On distribution lines, much of the voltage drop is due to a circuits inductance. If we add a capacitor in series with the inductance, the capacitor cancels out the inductance262. Now, the circuit has less total inductance, so fluctuations in load cause less voltage drop. A series capacitor is a passive circuit elementit responds instantaneously and automatically. Series capacitors help with voltage regulation as well as with improving voltage flicker. Electrically, series capacitors elegantly solve voltage flicker; in practice, they are not widely used, mainly because of reliability of short-circuit protection and cost. Historically, spark gaps were used to protect the capacitors during downstream faults. Utilities have had many problems with failures. Series capacitors are nonstandard and must be custom designed.

Capacitor Switching and Tripping of Adjustable-Speed Drives

Simultaneously switching all three phases of a capacitor bank results in a major transient event for the electrical system. It is characterized by a sudden removal of energy because of the zero voltage on the capacitor, followed by a high-frequency ringing while the inductive and capacitive elements cause a damped sine wave riding on the 60-Hz fundamental. The magnitude of this event is determined by the voltage across the capacitor at the instant of closure with the system and the available current from the system. The transient shown in Figure 10-5 is an example of a capacitor-switching transient; this is severe enough to have a high probability of disrupting customer equipment, as will be explained later.

S.A. Miske, Considerations for the Application of Series Capacitors to Radial Power Distribution Circuits, IEEE Transactions on Power Delivery, Vol. 16, No. 2, April 2001, pp. 306318.

262

10-10

Specific Power Quality Issues

Figure 10-5 Transient Caused by Switching of Capacitor at a Random Point of the Sine Wave

Transient frequencies due to utility capacitor switching usually fall in the range 300 to 1000 Hz.
f = 1 2 Ls C = 60 MVAsc MVAR
Eq. 10-3

Typically, switching surges range in magnitude from 1.1 to 1.6 pu. Switching surges can be magnified inside customer facilitiesthese transients can cause malfunctions in some types of sensitive equipment. Switching surges can also be magnified on certain distribution feeders. Magnification of the capacitor energization transient can occur when the series combination of a step-down transformer and a lower-voltage capacitor bank causes the energizing transient frequency to be magnified at the lower voltage capacitor. The conditions for magnification are:

C1 >> C2 (the upstream capacitor is much larger than the load-side capacitor). The natural frequency of high-voltage and low-voltage circuit is close. No significant resistive load in the low-voltage circuit.

Transient overvoltages resulting from capacitor switching are usually not a significant concern to the utility, because peak magnitudes are below the level at which utility surge protection, such as arresters, begins to operate. Because of the relatively low frequency, these transients will pass through step-down transformers to customer loads.

10-11

Specific Power Quality Issues

Unlike closing a capacitor bank, opening one should be an extremely mundane system event. The current interruption occurs at a current zero, and therefore system disturbance is negligible. Opening a capacitor bank should be mundane, but a severe transient can occur if the capacitor switch restrikes. As significant as the transients associated with random-point closing may appear, an even more disruptive transient occurs when an opening capacitor switching exhibits restrike. As previously mentioned, the capacitor current is interrupted at a current zero. This, of course, corresponds with the maximum voltage across the capacitor. As a result, the voltage across the interrupting contact is immediately equal to the phase-to-phase or phase-to-neutral voltage (depending on system connection). This voltage remains on the capacitor as a direct current (dc) value, only decaying slowing because of high-value internal discharge resistors. As soon as the contact parts, the system voltage begins to build in the polarity opposite to the trapped charge. This places extremely high requirements on the dielectric to become fully restored, even as the contacts continue to part. Whether because of mechanism binding or contact deterioration, it is not rare to have the dielectric break down as the system voltage approaches its opposing maximum. This effectively now re-energizes the capacitor at twice its rated voltage and creates even more severe transients. Actual restrikes have been observed on even the next peak voltage (two separate restrikes, same opening). Figure 10-6 illustrates restrike on a grounded-wye bank.

Figure 10-6 Illustration of a Restrike on a Grounded-Wye Bank

Effect on Customer Equipment Customer equipment affected by these transients reacts as peak detectorseven if buried within the primary power flow elements of other devices263. All devices that use a dc link for operation are susceptible to transients. The input for these devices involves a diode bridge, controlled or not, and a capacitor bank. Figure 10-7 shows a typical ac transient caused by capacitor
263

McGranaghan, M.F., Grebe, T.E., Hensley, G., Singh, T., and M. Samotyj, Impact of Utility Switched Capacitors on Customer Systems. II. Adjustable-Speed Drive Concerns," IEEE Transactions on Power Delivery, Vol. 6, No. 4, October 1991, pp. 16231628.

10-12

Specific Power Quality Issues

switching; the typical ASD topology; and, finally, the effect on the dc bus voltage. The dc bus voltage is shown for an ASD with and without a reactor, or choke, on the input. The improvement with the use of a relatively inexpensive and easily applied reactor is obvious.
800

600

Voltage (V)

400

200

-200

-400

-600 0 20 40 60 80 100

Time (mS)

DC Bus Rectifier Inverter Inverter

Induction Motor

760V 660V
DC Bus Voltage

trip level

Figure 10-7 Effect of Capacitor-Switching Transient on the Direct Current Bus of an Adjustable-Speed Drive

Although transients created by capacitor switching may cause drives to trip, ASD manufacturers also express concern regarding the cumulative effect of electrical component overstresses that may not manifest themselves until an outright failure occurs. One class of devices subject to overstress includes semiconductor power devices. Coincident with the dc voltage transient shown in Figure 10-7 is a current surge that may easily exceed the devices -cycle current-surge rating. Semiconductor fusing often uses fault-limiting devices that, by their design, are subject to metallurgical fatiguing when subjected to repeated mechanical and thermal cycling. These faultlimiting devices may open unexpectedly and unexplainably at a later time unrelated to a fault.

10-13

Specific Power Quality Issues

Solutions to Switching Transients Tracking down malfunctions of adjustable-speed drives (or other electronic equipment) can be tricky. Drives and other sensitive equipment can trip for a variety of reasonson the power system side, voltage sags or swells could trip a drive; the drive could also trip for load-side reasons or because of control problems. Power quality monitors installed at the drive are the best way to verify the power quality problem. If drive trips occur regularly and those times tend to coincide with typical closing times of capacitors (8AM), that is a likely sign of capacitorswitching problems. Once a problem is identified as a transient, finding the offending capacitor is usually straightforward. Two utility-side means of limiting the transient have been successfully employed:

Synchronous closing: Individual phase switching is common, as mentioned previously. This makes possible the synchronous closing of the switch. As stated earlier, synchronous closing refers to closing the contacts at a point on the sine wave of extremely little voltage, preferably zero. This option does not apply where all contacts are operated simultaneously, as in the case of most distribution-level circuit breakers and all circuit switchers. Pre-insertion impedance: The transient may also be reduced by closing the system through an impedance and limiting the current until the main contacts are closed. The capacitor is thus pre-charged. This is termed a pre-insertion resistor or pre-insertion inductor. This device has a limited capacity and is incorporated into the mechanism driving the main contacts. It is in effect shorted out by the main contacts during normal operation.

For line capacitors, synchronous closing is the only option. Station capacitor banks have options 264 for either solution. For line capacitors, a 2002 EPRI survey found that only 20% of utilities used synchronous switches, and of those, most were either on a trial basis or for targeted areas. The newest synchronous switching schemes employ active feedback to compensate for physical changes and maintain precise position control by the use of voice coil actuators. The operation in the synchronous closing mode requires that voltage from the source be available. The actual transient produced upon the synchronous closing of a capacitor bank at a typical installation is shown in Figure 10-8.

Improved Reliability of Switched Capacitor Banks and Capacitor Technology, EPRI, Palo Alto, CA: 2002. 1001691.

264

10-14

Specific Power Quality Issues

Figure 10-8 Transient Caused by Synchronous Switching of a Capacitor Bank

The best local solution is the addition of a series inductor (line reactor) connected to the input terminals to the adjustable-speed drive. Other inexpensive utility-side options are to use clock controls and switch the capacitor on (for most capacitor switches, the switch on is the most likely to cause a transient) during process off times (such as before 5 AM). Another option is to disable switching on a capacitor unit. This may increase distribution line losses, but as long as circuit voltages are still okay, service to other customers is not degraded. One could also investigate whether the problems are caused by resonances that magnify transients. If so, disconnecting one of the capacitors or changing its size will reduce the severity of the transient, possibly enough to stop malfunctions of the drives.

10-15

Specific Power Quality Issues

Harmonics
Harmonics are sinusoidal voltages or currents having frequencies that are integer multiples of the frequency at which the supply system is designed to operate (termed the fundamental frequency, which is usually 50 Hz or 60 Hz). Harmonic distortion exists due to nonlinear characteristics of the power system and its connected loads. Most harmonic distortion originates with nonlinear devices on the power system. Nonlinear devices produce non-sinusoidal current waveforms when energized with a sinusoidal voltage. Examples of these devices are adjustable-speed drives, switch-mode power supplies (including computers and other office equipment), fluorescent lighting, battery chargers, transformers, generators, and arc furnaces. Although power system components such as transformers and generators contribute to some distortion, the bulk of the distortion in most power systems comes from nonlinear end-use devices. These devices can usually be modeled as current sources that inject harmonic currents into the power system. Voltage distortion results because these currents cause nonlinear voltage drops across the system impedance. Harmonic distortion is a growing concern for many customers and for the overall power system due to the increasing application of power electronics in end-use equipment. Harmonic distortion levels can be characterized by the complete harmonic spectrum with magnitudes and phase angles of each individual harmonic component. It is also common to use a single quantity, the total harmonic distortion, as a measure of the magnitude of harmonic distortion. For example, the total current harmonic distortion is:

THD

I
h=2

2 h

I1

Eq. 10-4

where Ih is the RMS current of the hth harmonic current, and I1 is the RMS value of the fundamental current. A typical voltage waveform does not exceed 5% THD. However, the power supply input current THD could exceed 100%. Harmonic currents result from the normal operation of nonlinear devices on the power system. Figure 10-9 illustrates the waveform and harmonic spectrum for a typical adjustable-speed drive input current. Current distortion levels can be characterized by a total harmonic distortion, as described above, but this can often be misleading. For instance, many adjustable-speed drives will exhibit high total harmonic distortion values for the input current when they are operating at very light loads. This is not a significant concern, because the magnitude of harmonic current is low, even though its relative distortion is high.

10-16

Specific Power Quality Issues

T Y P E 1 W aveform 1 00 H P P W M A SD - N o C hoke
400
400

T Y P E 2 W aveform 100 H P P W M A SD - 3% C hoke

200

200

Am ps
-200

A m ps
-200

-400 0.05 0.06 0.07 0.08 0.09 0.10

-400 0.05 0.06 0.07 0.08 0.09 0.10

Time (Seconds)
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1 3 5 7 9 11 13 15 17 19 21 23 25 100%

Time (Seconds)

I T H D = 8 0.6% I R M S = 1 48 .2 A m p s I F u n d = 1 15 .4 A m p s

90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1 3 5 7 9 11

I T H D = 3 7.7% I R M S = 1 17 .6 A m p s I F u n d = 1 10 .1 A m p s

Figure 10-9 Waveform and Harmonic Spectrum of Typical 6-Pulse ac Motor Drive

To account for the relative harmonic current levels in a consistent fashion, IEEE Std. 519-1992 defines another term, total demand distortion265 . Total demand distortion is the same as the total harmonic distortion, except that the distortion is expressed as a percent of some rated load current rather than as a percent of the fundamental current magnitude. Recommended practice for acceptable harmonic current and voltage distortion levels on distribution and transmission circuits are provided in IEEE Std. 519. The increased voltage distortion resulting from a relatively weak source such as a generator can 266 cause potential problems for customer loads. The main issues are as follows :

13

15

17

19

21

23

25

Ha rm on ic Nu mb er

Ha rm on ic Nu m be r

The voltage-sensing circuit in an uninterruptible power supply that determines the transfer from utility line to battery during an undervoltage condition may be susceptible to a distorted voltage waveform, causing the UPS to go to battery power. Its batteries may be depleted after only a short time during the operation of the DG. Increased voltage distortion will reduce the life of induction motors and other customer loads. A building power-line carrier signal, clock-synchronizing signal, or other communication signal within the customer premises may be susceptible to increased voltage distortion. Equipment that uses a front-end filter with capacitors such as lighting ballasts may be damaged due to increased voltage distortion.

McGranaghan, M.F., Grebe, T.E., Hensley, G., Singh, T., and M. Samotyj, Impact of Utility Switched Capacitors on Customer Systems. II. Adjustable-Speed Drive Concerns," IEEE Transactions on Power Delivery, Vol. 6, No. 4, October 1991, pp. 16231628. 266 IEEE Task Force, Effects of Harmonics on Equipment, IEEE Transactions on Power Delivery, Vol. 8, No. 2, April 1993, pp. 672680.

265

10-17

Specific Power Quality Issues

Power-factor-correction capacitors within the customer premises may also be damaged due to excessive harmonic current due to increased voltage distortion.

When harmonics cause problems, capacitors are often a contributing factor. Capacitors can cause resonances that amplify harmonics. This often shows up first at the capacitor bank, where harmonics cause fuse operation or even capacitor failure. Most problems with severe harmonics are found in industrial facilities, where capacitors resonate against the system impedance. Utility resonances are rare but do sometimes cause problems. One capacitor on a distribution circuit will resonate against the inductance back to the system source (including the line impedances and transformer impedances), as shown in Figure 10-10.
Impedance upstream of the capacitor X1 X2 X3

XC Harmonic load injected along the circuit Capacitor

High voltage distortion

Figure 10-10 Harmonic Resonance

The resonance point between the capacitor and the system is (this is the same frequency that the system will ring at during a switching surge):
n= XC = XS MVAsc MVAR
Eq. 10-5

where n XC XL = order of the harmonic, = line-to-ground impedance of one phase of the capacitor bank at nominal frequency, = system impedance at nominal frequency,

MVASC = three-phase short-circuit MVA at the point where the capacitor is applied, and MVAR = three-phase Mvar rating of the capacitor. If a nonlinear load is injecting a harmonic frequency equal to the systems resonance point, the circuit can have overvoltages. Common danger frequencies are n=5, 7, 11, and 13. Larger capacitors lower the resonant point to where it is more likely to cause problems. Multiple 10-18

Specific Power Quality Issues

capacitors on a circuit create multiple resonant points that can require more sophisticated analysis. The worst conditions are when the harmonic source is right at the capacitor or downstream of the capacitor. Harmonics injected further upstream are amplified less. Solutions to Harmonics Identifying harmonics is easy to do with an oscilloscope-type meter that measures voltage and current. Tracing the source of the harmonics is more difficult, especially if a circuit has capacitors. Before chasing after the harmonics, make sure that the problem (capacitor fuse operations, equipment operations, and so on) really is caused by the harmonics. If it is, the utility-side options are harmonic filters or distribution system changes. Before resorting to filters, other utility-side options could be explored. Many harmonic problems originate because of resonances with capacitors, so we can often greatly reduce problems by disrupting the resonance. To disrupt the resonance, move capacitors or change their size or switch them off. Usually, it does not take much of a change to move a resonant point enough to reduce the harmonic amplification. These are easy but may require some trial-and-error. While not a mainstream or regular application, several manufacturers provide tuned filter banks that can be used to filter objectionable harmonics on distribution circuits. These are available in pole-mounted or pad-mounted configurations. Most of the applications are for industrial use, but some are used on utility distribution circuits. Harmonic filter banks consist of one or more series L-C tuned filters, as shown in Figure 10-11. Filters are normally tuned to just below the offending harmonic.

5th
Figure 10-11 Tuned Harmonic Filter

7th

Filters accomplish two objectivespower-factor correction and shunting one or more harmonic currents to ground. A series-tuned filter is constructed in each phase to ground by placing a choke in series with a shunt capacitor and then tuning the choke so that the inductive and capacitive reactances are equal but opposite at the desired harmonic. Normally, a filter bank is custom-engineered for a given problem. Harmonic filters must be properly designed. Some key issues are:

Tuning: Tuning a filter slightly below the desired harmonic, for example at the 4.7th instead of the 5th harmonic, helps to reduce capacitor voltage without significantly degrading filter performance. 10-19

Specific Power Quality Issues

Sizing: Care must be taken to dedicate enough kvar to the filter. In most cases, the filter kvar should be approximately the amount needed for power-factor correction. Filters with smaller kvar will have sharp tuning curves and will be easily overloaded by stray harmonics that are present in the network. Voltage rating: Because a filter capacitor usually experiences 1.2 to 1.3 p.u. rms voltage, plus significant harmonics, care must be taken that the capacitor voltage rating is adequate. The fact that kvars decrease by the square of voltage must also be taken into consideration.

Interaction: Filters improperly applied to existing capacitor banks can cause problems. A nontuned capacitor adjacent to a filter will resonate with the filter at some frequency. If that th frequency corresponds to a key frequency (like the 7 harmonic), harmonics may be worse than without the filter. Proper application of harmonic filters with capacitors nearby requires a careful engineering study to ensure that there are no improper interactions between filters and capacitors. Separately designed filters also can interact with each other. This is a key constraint for distribution applicationsover the years, adjacent capacitors could be installed on the same circuit without realizing that there are tuned filters on the circuit.

Telephone Interference
Telephone interference has been a harmonics-related concern for many decades, but the gradual phasing out of open-wire telephone circuits has reduced the number of interference problems. But, when interference problems occur, they can be difficult to trace and to fix. While the frequency response of the combined telephone circuit and human ear is largely immune to 60-Hz interference, higher harmonics fall into the low-audio range. When harmonic currents on power lines inductively couple into nearby phone lines, they can cause significant interference. Typically, the problem harmonics are either characteristics sixpulse harmonics due to large converters or 9th and higher multiples of three (that is, zerosequence) due to transformer saturation. All things being equal, zero-sequence harmonics are more problematic than positive- and negative-sequence harmonics because a-b-c zero-sequence fields are additive and therefore do not decay as rapidly with distance. The telephone influence factor (TIF) curve shown in Figure 10-12 gives the relative interference weighting that applies to inductively coupled harmonic currents flowing in power lines. The frequencies from 1000 to 3800 Hz are the most criticalthese are most likely to cause interference.

10-20

Specific Power Quality Issues

12000 10000 8000 TIF 6000 4000 2000 0 0 600 1200 1800 2400 3000 3600 4200 Fre que ncy - Hz

Figure 10-12 Telephone Influence Factor (TIF) Curve

Solutions to Telephone Interference Telephone interference problems are usually solved by the telephone company in cooperation with the electric utility involved. Solutions are often trial-and-error. The normal solutions 267,268,269 : are

Change size: For problems involving a resonance, increasing or decreasing the size of a bank can shift the resonance point enough to ease interference problems. The easiest solution is to disconnect the bank that is contributing to the problem. This is also a good first step to quantify the role of the capacitor bank in the interference. Move the bank: Moving a capacitor can change a resonance point enough to stop interference problems. On some circuits, one can also move the capacitor away from the telephone circuits having problems. Unground the bank: A floating-wye connection has no connection to ground, so the connection blocks zero-sequence harmonics. Two-bushing capacitor units are necessary for floating the wye point (unless the utility floats the capacitor tanks and deals with the safety issues that accompany that practice). Add a neutral filter: While not a common solution, a tuned reactor in the ground path of a wye-connected capacitor bank is invisible to positive and negative sequences, but it changes the zero-sequence resonant frequency of a distribution feeder and often eliminates the resonance problem.

IEEE Std. 1137-1991, IEEE Guide for the Implementation of Inductive Coordination Mitigation Techniques and Application. 268 IEEE Std. 776-1992, IEEE Recommended Practice for Inductive Coordination of Electric Supply and Communication Lines. 269 IEEE Working Group, Power Line Harmonic Effects on Communication Line Interference, IEEE Transactions on Power Apparatus and Systems, Vol. PAS-104, No. 9, September, 1985, pp. 25782587.

267

10-21

Specific Power Quality Issues

The frequency of the noise signal can often help with identifying the source. If the signal is the third or ninth harmonic (as is often the case), then excessive zero-sequence current is the problem, possibly influenced by a resonance. Other frequencies, such as the fifth or seventh, point to a typical three-phase converter. If one of the frequencies is much higher than the others, look for a resonance causing amplification.

10-22

11
SUMMARY
Todays utility customers demand high reliability and improved power quality from their electric power supply. The primary driver behind this trend is the increasingly sensitive nature of customer loads. As we have ushered in the digital age, sensitive electronic loads have spread from industrial facilities to commercial customers and finally to widespread implementation in residential homes. The seemingly ubiquitous microchip has embedded itself in everything from multi-million dollar production facilities to inexpensive home appliances. The result of these advancements is that customers can no longer tolerate the type of power quality events that would go unnoticed only a few years ago. In essence, the benchmark of what constitutes acceptable power quality is changing, as customer loads become more sophisticated and more sensitive to disturbances. The increasingly stringent power supply needs of customers are forcing modern electric utilities to re-think the design and construction of electric distribution circuits. The three mostsignificant power quality concerns for most customers are:

Voltage sags Momentary interruptions Sustained interruptions

Power quality problems affect different customers differently. Most residential customers are affected by sustained and momentary interruptions (also known as momentaries). For commercial and industrial customers, sags and momentaries are the most common problems. Each circuit is different, and each customer responds differently to power quality disturbances. Voltage sags, momentaries, and sustained interruptions are caused by faults on the utility power system, with most of them on the distribution system. Because faults are the root cause of voltage sags and interruptions, reducing the number of faults will obviously reduce the number of voltage sags and interruptions and power quality will improve. Although faults can never be totally eliminated, there are different methods available to minimize the impact on customers. Table 11-1 shows a summary of the techniques discussed in this report. These techniques range in complexity from simply reviewing fuse sizing standards to more involved approaches for reducing the number of faults occurring on the system.

11-1

Table 11-1 Summary of Utility-Side Options for Improving Power Quality Within the Distribution System Voltage Sags

Use fuse saving Use current-limiting fuses Use smaller lateral fuses Use faster breakers or reclosers Raise the nominal voltage. Reduce faults.
Momentary Interruptions

Use an instantaneous reclose Use fuse blowing Use single-phase reclosers Install extra downstream devices: fuses or reclosers. Use sequence coordination with downstream devices Reduce faults.
Sustained Interruptions

Use fuse saving Install extra downstream devices: fuses, reclosers, or sectionalizers. Use faulted-circuit indicators. Reduce faults.

11-2

Export Control Restrictions Access to and use of EPRI Intellectual Property is granted with the specific understanding and requirement that responsibility for ensuring full compliance with all applicable U.S. and foreign export laws and regulations is being undertaken by you and your company. This includes an obligation to ensure that any individual receiving access hereunder who is not a U.S. citizen or permanent U.S. resident is permitted access under applicable U.S. and foreign export laws and regulations. In the event you are uncertain whether you or your company may lawfully obtain access to this EPRI Intellectual Property, you acknowledge that it is your obligation to consult with your companys legal counsel to determine whether this access is lawful. Although EPRI may make available on a case-by-case basis an informal assessment of the applicable U.S. export classification for specific EPRI Intellectual Property, you and your company acknowledge that this assessment is solely for informational purposes and not for reliance purposes. You and your company acknowledge that it is still the obligation of you and your company to make your own assessment of the applicable U.S. export classification and ensure compliance accordingly. You and your company understand and acknowledge your obligations to make a prompt report to EPRI and the appropriate authorities regarding any access to or use of EPRI Intellectual Property hereunder that may be in violation of applicable U.S. or foreign export laws or regulations.

The Electric Power Research Institute (EPRI) The Electric Power Research Institute (EPRI), with major locations in Palo Alto, California, and Charlotte, North Carolina, was established in 1973 as an independent, nonprofit center for public interest energy and environmental research. EPRI brings together members, participants, the Institutes scientists and engineers, and other leading experts to work collaboratively on solutions to the challenges of electric power. These solutions span nearly every area of electricity generation, delivery, and use, including health, safety, and environment. EPRIs members represent over 90% of the electricity generated in the United States. International participation represents nearly 15% of EPRIs total research, development, and demonstration program.

Together...Shaping the Future of Electricity

Program:
2006 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc.
Printed on recycled paper in the United States of America

Power Quality Analysis Tools and Testing

1010192

ELECTRIC POWER RESEARCH INSTITUTE

3420 Hillview Avenue, Palo Alto, California 94304-1395 PO Box 10412, Palo Alto, California 94303-0813 USA 800.313.3774 650.855.2121 askepri@epri.com www.epri.com