Best Practices in Vegetation Management. For Enhancing Electric Service in Texas

Best Practices in Vegetation Management For Enhancing Electric Service in Texas PUCT Project 38257 Submitted to: Submitted by: Public Utility Commission of Texas 1701 N. Congress Avenue Austin, TX 78711-3326 Texas Engineering Experiment Station TEES project B2220 Texas A&M University System College Station, Texas 77843 Date: November 11, 2011 Principal Investigator: Contact Information: B. Don Russell, Ph.D., P.E. Dept. of Electrical and Computer Engineering Texas A&M University Co-Investigators: College Station, TX 77843 Jeff Wischkaemper Phone: 979-845-7912 Carl Benner Email: bdrussell@tamu.edu 1

Project Disclaimer: The opinions and conclusions set forth in this report are solely those of the principal investigator and do not represent an official position of the Texas Engineering Experiment Station, Texas A&M University System, or the Public Utility Commission of Texas. The primary and overriding emphasis during the investigation was the relationship between vegetation practices and reliability as measured by service continuity metrics. Other vegetation issues including public safety, fire prevention, and aesthetic considerations were not directly investigated but are mentioned, as appropriate, in the text in relation to reliability. All results, conclusions, and opinions expressed herein should be used carefully within the context and assumptions set forth in the report. The conclusions stated are subject to change at any time based on further research investigation or the acquisition of new data leading to different scientific conclusions. The best practices of the scientific method have been utilized in all analyses and conclusions set forth in this report. Conclusions and findings have been stated to a reasonable degree of engineering certainty; exceptions may exist due to specific conditions. This investigation was not an exhaustive review of all vegetation management practices and related literature but does document common practices and their supporting science. This project has drawn from the work of many professionals in the electric utility and vegetation industry including arborists, electrical engineers, and researchers. Reference has been made to these individuals or their work or publications within the document and/or bibliography, as appropriate. If any work has been used, referenced, quoted or borrowed without proper attribution, we apologize; this was not our intent. 2

Acknowledgements and Recognition This work has drawn heavily on the work of various individuals and groups, including the following. The Electric Power Research Institute which has sponsored numerous research projects in vegetation management. Thomas A. Short, author of Distribution Reliability Power Quality, Taylor & Francis. John W. Goodfellow, whose research and experimentation and writings in vegetation management are highly respected and with whom insightful discussions were held. Researchers with the Power System Automation Laboratory of Texas A&M University, who have performed staged vegetation faults and whose research has captured and archived the best recorded examples of naturally occurring vegetation faults and their effects on the electric distribution system. Vegetation management experts and professionals in utilities in Texas who freely discussed their vegetation practices but, for purposes of confidentiality and anonymity, will remain unnamed. Others who provided valuable assistance include the following individuals. Jeff Wischkaemper who, based on research he conducted, co-authored sections of the report on the mechanisms of vegetation-caused arcing faults. Carl Benner, with whom valuable discussions were held concerning practical and cost effective VM practices. Sharon Loe, who provided clerical support and formatting. Jessica Meadors, who assisted with clerical support and procurement of referenced documents. 3

Executive Summary Vegetation intruding into overhead power lines has produced significant negative impacts on reliability since the earliest days of electricity distribution. An electric utility must address vegetation management comprehensively to maintain reliability of electric service to customers. Expenditures for vegetation management typically represent one of the largest recurring maintenance expenses for electric utilities. Vegetation-related outages have two fundamental underlying mechanisms: mechanical and electrical. The mechanical mechanism occurs when, for example, a proximate tree falls into a distribution line, breaking conductors or other line apparatus. The electrical mechanism occurs when vegetation comes into contact with intact, energized conductors, and the resulting unintended flow of electrical current either causes localized erosion and weakening of conductors, or a high-current flashover that trips system protection. Of the two mechanisms, mechanical tear-down causes the majority of vegetationcaused outages. Identification and remediation of hazard trees are therefore key elements of an effective vegetation-management program. The electrical mechanism causes a lesser number of vegetation-related outages. Single-phase lines on lower-voltage distribution systems (e.g., 15kV class, the predominant distribution voltage in Texas and the United States) typically have insufficient voltage to cause significant electrical activity when touching vegetation. Therefore, the electrical mechanism predominantly affects only higher-voltage distribution systems or three-phase feeder sections where phase-to-phase voltage is available. Reliability indices such as System Average Interruption Duration Index (SAIDI) and System Average Interruption Frequency Index (SAIFI) are valuable means to track reliability. To quantify the effectiveness of vegetation-management programs, however, it is necessary to separately calculate these indices based only on vegetation-caused outages and interruptions. These vegetation-specific indices can only be calculated and properly evaluated if detailed information is collected on the cause of feeder outages, including specific data concerning the vegetation-related cause(s) of an event. Given that mechanical tear-down, often from off right-of-way trees and limbs, represents the majority of vegetation-caused outages, trimming practices that attempt to keep small branches a few feet from utility conductors will have a relatively minor effect on system reliability. By contrast, trimming practices that seek to eliminate tear-down conditions or multi-phase faults on three-phase feeder sections will yield a significantly greater improvement to overall reliability. This is particularly true during ice, snow, and windstorm conditions, when falling trees and broken overhead branches result in substantial numbers of both electrical events and mechanical tear-downs. Trimming right-of-way vegetation on a fixed time cycle (e.g. every three years) will seldom achieve maximum reliability at optimal cost when applied uniformly across an entire utility. Rigid adherence to a system-wide fixed cycle for trimming of vegetation near conductors, without respect to local conditions, does not directly address the primary cause of outages, namely tear-down from right-of-way and off right-of-way trees, and blown and/or falling limbs. Furthermore, fixed trim cycles risk focusing too much attention to areas that have a good reliability history and too little attention on areas needing critical, more timely action. However, a targeted fixed-trim period, based on past feeder performance, 4

may be appropriate if adjustments are allowed based on annual variations and diversity in local conditions. Based on the last decade of study of vegetation management and arborist practices, and using the best science of condition-based maintenance protocols, this report recommends a reliability-centered program. In such programs, heavy emphasis is placed on inspection and condition-based decision making by vegetation professionals using continuously updated data on vegetation-caused outages. A reliability-centered program allows a utility to choose practices that can be selectively applied based on variations across the utility s service area. Factors such as annual growth rates, tree species, feeder construction type and feeder voltage can be taken into account to achieve optimal reliability for a given expenditure of funds. For example, drought simultaneously increases tree mortality and fire danger, but conversely slows vegetation growth into lines. Therefore, extended periods of low rainfall may require changes in VM practices. A reliability-centered vegetation program must engage the public, so that it understands the necessity of vegetation management. Success of a reliability-centered program requires continuity of vegetationmanagement expenditures to enable proper planning over multiple years. 5

Table of Contents Executive Summary Table of Contents Glossary of Terms 1.0 Project Methodology 2.0 Vegetation Management Introduction 2.1 Overview 2.2 Areas of Emphasis 3.0 The Nature and Mechanism of Vegetation-caused Outages 3.1 Typical Questions and Issues 3.2 Outage Causation 3.3 Mechanical Tear-Down 3.4 Electrical Causes of Faults 3.5 Fire and Electrical Injury Hazards 3.6 Learning Points 4.0 Case Studies Experience of Utilities Documenting Vegetation Management Effectiveness 4.1 Example One Seattle City Light 4.2 Example Two Oncor Snow Storm 4.3 Learning Points 5.0 Reliability Indices and Vegetation Management Metrics 5.1 Metrics Defined 6

5.2 SAIFI/SAIDI Accuracy and Sensitivity 5.3 Vegetation Specific SAIDI/SAIFI Measures of VM Success 5.4 Documentation of Vegetation-caused Outages Data Collection and Accuracy 5.5 Learning Points 6.0 Optimal Reliability Improvement by Targeting Resources 6.1 Targeting Three-phase Main Feeder Sections 6.2 Trimming Frequency Versus Trimming Criteria 6.3 Mandatory Clearance Requirements 6.4 Cost-Benefit Analysis and Resource Prioritization 6.5 Learning Points 7.0 National Standards 7.1 Occupational Safety and Health Administration (OSHA) 7.2 National Electrical Safety Code (NESC) 7.3 American National Standards Institute (ANSI) 7.4 Federal Standards 7.5 Learning Points 8.0 Vegetation Management Scheduling Practices - Overview and Discussion 8.1 Periodic Fixed Cycles and Cycle Length 8.2 Outage Response Clearing 8.3 Reliability Centered, Condition Based Scheduling 8.4 Hazard/Danger Tree Identification and Remediation 8.5 Evaluation of Common Scheduling Practices 7

9.0 Vegetation Outage Assessment, Data Collection, and Reporting Programs 9.1 Documenting Vegetation Outages 9.2 Calculation of Vegetation Specific SAIDI/SAIFI Indices 10.0 Other Considerations 10.1 Public Education and Interaction 10.2 Species Specific/Seasonal Specific Practices 10.3 Clearance Requirements during Scheduled Trimming 10.4 Constancy in VM Budgets 11.0 Recommendations 11.1 Applicable ANSI, NESC, and OSHA Provisions Should be Adopted 11.2 Vegetation Specific SAIFI/SAIDI Indices should be Calculated and Reported on an Annual Basis 11.3 Mandatory Minimum Clearance Requirements Should Not Be Adopted 11.4 Reactive Vegetation Clearing in Response to Outages is Appropriate, but Insufficient 11.5 System Wide Fixed Trim Cycles are not Recommended as the Sole Uniform Practice 11.6 A Reliability Centered Vegetation Program (RCVP) Should be Designed and Adopted by each Utility 11.7 Public Awareness and Engagement Programs are of Critical Importance 11.8 Proactive Programs can Reduce Future Vegetation Management Costs and Should be Encouraged 11.9 A Statewide Vegetation Management Public Relations Discussion and Consensus Are Needed 11.10 Future Areas of Investigation and Opportunity 8

List of Figures in Main Text List of Figures in Appendix A Appendices Appendix A: How Electrical Faults Occur as a Result of Vegetation Intrusion Appendix B: Naturally Occurring Vegetation Outage Case Studies Appendix C: PUCT Workshop Agenda Appendix D: Bibliography and Suggested Reading References 9

Glossary of Terms Burn-down A burn-down results when an arcing fault generates enough conductor or apparatus damage to result in a conductor breaking and/or falling to the ground. This may occur because of the conductor melting in two, or through burning down of a utility pole. Fault A fault is an unintended, abnormal flow of electrical current within a circuit. In a power system, a fault occurs when a conductive path is formed between an energized conductor and another phase conductor, the system neutral, or ground. Interruption As per IEEE 1366, an interruption is the loss of service to one or more customers connected to the distribution portion of the system. It is the result of one or more component outages, depending on system configuration. Lockout The permanent opening of a substation recloser, generally occurring after two or three reclose attempts. Momentary interruption As per IEEE 1366, a momentary interruption is a single operation of an interrupting device that results in a voltage zero. For example, two circuit breaker or recloser operations (each operation being an open followed by a close) that momentarily interrupts service to one or more customers is defined as two momentary interruptions. Outage As per IEEE 1366, an outage is the state of a component when it is not available to perform its intended function due to some event directly associated with that component. For the purposes of this report, the term outage is used to describe a sustained interruption of power to customers. Recloser A recloser is a protective device which attempts to clear a transient fault by removing power from the system for a short period of time. This sequence is generally referred to as a trip and reclose, and may recur several times. Approximately 90% of distribution faults are temporary, and can be cleared by reclose operations. Note that each recloser operation results in a momentary interruption. SAIDI System Average Interruption Duration Index. This number represents the number of minutes an average customer would expect to be without power during a year. SAIFI System Average Interruption Frequency Index. This number represents the number of outages an average customer would expect during a year. Single-phase lateral A single-phase lateral is a portion of a radial distribution circuit that branches off from a three-phase section, consisting of a single-phase conductor and neutral. A fault on a singlephase lateral does not typically result in the interruption of power to the three-phase section it is attached to. 10

Tear-down A tear-down occurs when mechanical forces result in the destruction of a power line or other system apparatus; for example, a tree falling and breaking a power line, or a car snapping a utility pole in half as a result of a traffic accident. Any action that breaks mechanical supporting insulators, rips conductors from poles, or breaks and drops conductors can generally be referred to as mechanical tear-down. Three-phase feeder section A three-phase section of line is the portion of a distribution circuit consisting of three-phase conductors and a system neutral. A fault on a three-phase section of line will result in the interruption of power to all three-phase and single-phase sections of line downstream of the faulted section. Three-phase tripping When a single-phase fault occurs on a three-phase section of line, the practice of three-phase tripping interrupts power to all three phases, instead of interrupting power to only the faulted phase. Vegetation management (VM) Vegetation management includes all planning, budgeting, inspection, pruning, tree removal and similar activities to control vegetation intrusion into power lines and apparatus. Voltage gradient A voltage gradient is the voltage spread across a distance. For example, 7,200V applied across 3 feet would result in a voltage gradient of 2,400 volts per foot. 11

1.0 Project Methodology The objective of this report is to document the known common practices of vegetation management (VM) in the context of the best science on this subject and to determine how these practices should be applied by Texas utilities. The intent is to recommend a framework for best practices in vegetation management given the geography, flora, weather, and typical infrastructure conditions of utilities in Texas. The VM practices of electric utilities across the United States, including many Texas utilities, have been carefully studied with specific emphasis on application in the state of Texas to improve distribution reliability. An attempt has been made to identify and formulate existing best practices; however, common practices and variations thereof have been noted throughout the report. Economic issues were considered only in a relative fashion to compare practices and discuss budget optimization. Vegetation standards commonly used by utilities were considered; however, this project did not address the procedures of arborists for actual tree trimming, pruning, or vegetation treatment, except as these directly affect outage frequency. There was no experimentation or field work performed under this project. However, the prior experimentation of multiple researchers including researchers of the Power System Automation Laboratory of the Department of Electrical and Computer Engineering of Texas A&M University has been utilized as a basis for certain conclusions found in this report. Scientific literature, industry reports and surveys, and industry publications available as of the date of this report have been utilized to establish the common practices of vegetation management used in the electric utility industry. This project has concentrated on vegetation management practices for distribution systems at all common distribution voltages. This document does not directly address vegetation management practices for transmission right-of-way. This report is not an exhaustive study or analysis of all literature or statistical data available on vegetation management but rather cites such work only as required to support stated conclusions. A representative sample of peer reviewed and industry literature that provides the scientific basis for the conclusions of this report are cited and/or are listed in the bibliography to be used by the reader for further study. The reader will note that the phrase best practice is often used in quotes. This is intentional to emphasize the following points. Delineating best practices is inherently problematic, if not subjective. This is due to the wide variation in conditions and circumstances that exist across utilities. What is best for one is of secondary importance to another. Is a practice considered best because it reduces cost, or because it produces greater reliability? Is it best because it is in all cases safer to the public? One could define best practices as those that achieve the highest technical performance and safety at the lowest cost; but in the minds of some, cost should never be a criterion for best. One thing is clear as stated by C. S. Lewis, I am to give my readers not the best absolutely but the best I have. Defining best must necessarily be fluid, since the conditions at any time (e.g. rainfall) or place (e.g. forest versus prairie) may vary geographically or year-to-year across a given utility. This means that the best practices of this year may not best apply to next year. In a true sense there is no single set of best practices, but a shopping list of proven, science-based methods and procedures from which to design a good vegetation management program. It is on this basis that this project has proceeded. 12

2.0 Vegetation Management Introduction 2.1 Overview The utility industry has long recognized that vegetation is a significant cause of faults and outages in electric distribution systems. A fault occurs when a conductive path is formed between an energized conductor and another phase conductor, the system neutral, or ground. This may occur through direct contact, such as when two conductors are pushed together, or through an intermediary, for example when a tree branch spans two phase conductors for an extended period of time. Faults may result in momentary interruptions while protective devices attempt to automatically clear the faulted condition through a process known as reclosing, and may result in a sustained outage if the recloser cannot clear the fault. If the system is protected by fuses, an outage may occur when a fuse blows to clear a fault. Restoration of service requires manual replacement of the fuse. An outage also occurs when conductors are torn down by falling trees or vegetation. While statistics vary widely across utilities based on such factors as growth rates and the types of vegetation, it is generally thought that approximately 20% of distribution faults are related to or caused by vegetation. It is also generally understood that there is a strong correlation between weather events and vegetation-caused outages, and that there is a substantial increase in the duration of outages when they are caused by trees tearing down distribution infrastructure. Vegetation management or tree trimming has long been an emphasis of operating utilities and represents a major annual cost. It is commonly stated that vegetation management is the single largest maintenance expense for many utilities. [1] With the current emphasis on improving reliability, reducing the frequency and duration of outages, and improving storm recovery performance, utilities must be diligent in evaluating and improving vegetation management practices. This project has a wide range of purposes and objectives. The report seeks to identify best practices in vegetation management that can be employed by utilities in the state of Texas to improve reliability and distribution feeder performance under both normal and adverse weather conditions. The project seeks a balance between known science as to what can be achieved in vegetation management, versus practices that are reasonable, cost effective, and suitable (best) for electric utilities. This project has endeavored to understand the validity and uses of various metrics that attempt to measure reliability, including the unreliability caused by vegetation intrusion. The sensitivity and accuracy of these metrics and their value for quantifying the success of vegetation management procedures is presented. There has been no attempt to comprehensively document all vegetation management practices by all utilities in the United States. Major and pervasive practices have been included. The studies and research of selected utilities have been utilized, primarily as reported in the general literature; sources are hereafter cited. In general, it can be stated that the vegetation management practices of utilities in 13

the state of Texas are a representative sample of utility practices throughout North America. Furthermore, there is no significant vegetation management practice that is not practiced in some form by one or more utilities in the state of Texas, with the possible exception of maintaining continuous minimum clearance distances from lines for all vegetation. Comparisons of the practices of utilities in and outside Texas will be made throughout this document, as appropriate, to support the findings of this report. There is a broad misunderstanding on the part of the public and among some utility and vegetation management professionals as to the nature of electrical faults caused by vegetation. This is understandable since most vegetation professionals are not trained in electrical science. To avoid misunderstandings leading to incorrect conclusions, we have included a section in the report that discusses, in detail, the mechanisms of electrical fault creation through vegetation. It is hoped that this discussion will be of general use by vegetation professionals and electric power engineers. 2.2 Areas of Emphasis A review of vegetation management history, literature and practices, as well as recent research, reveals specific areas that must be emphasized when investigating VM best practices. Consider the following statements: Trimming and clearing of trees makes electricity delivery more reliable, as measured by: - Fewer outages - Shorter outages - Fewer momentary interruptions - Improved power quality Trimming and eliminating trees makes the distribution system safer. - Tear-downs of energized lines and equipment are reduced. - Burn down of conductors is reduced. Trimmed feeders reduce the time and cost of storm restoration. - Cleared rights-of-way are accessible to repair crews - Rights-of-way without vegetation debris to be removed speeds repairs. Vegetation management is expensive. - Often cited as the single highest recurring maintenance expense. The performance and effectiveness of vegetation management programs are difficult to evaluate on a short time scale. - Deferral of trimming may not immediately affect outage frequency. - Increased future costs due to poor VM practices today are difficult to document. 14

The above listed truisms were carefully investigated in this project since they provide the basis for understanding and evaluating vegetation control methods. These statements provide insight and context into the best practices for vegetation management. Specific emphasis is also given to how reliability can be quantified, documented, and tracked, as it is affected by vegetation management practices. The sensitivity of standard reliability indices to various VM practices is a key factor in deciding which practices should be labeled best and, therefore, should be adopted by utilities. Attention is also given to other negative consequences of vegetation intrusion on power lines. These include public safety considerations such as fires ignited by arcing from vegetation-caused downed power lines. 15

3.0 The Nature and Mechanism of Vegetation-caused Outages 3.1 Typical Questions and Issues Q1. How does vegetation cause outages on electric distribution feeders? Q2. What percentage of vegetation related outages are due to mechanical forces versus electrical short circuit conditions? Q3. What are the electrical characteristics and behavior of vegetation faults? Q4. What factors affect or exacerbate vegetation-caused outages? Q5. What is the contribution of hazard and danger trees as causes of distribution system outages? 3.2 Outage Causation Bare overhead conductors carry electricity at distribution voltage levels supported by insulators on poles. For most of the span above ground, air provides the insulation between energized conductors and/or neutral conductors or ground. Two mechanisms exist to compromise overhead, air insulated conductors. Any action that breaks mechanical supporting insulators, rips conductors from poles, or breaks and drops conductors can generally be referred to as mechanical tear-down. The consequences of mechanical tear-down include falling energized conductors. Another mechanism which compromises overhead conductors is the creation of short circuit conditions caused by objects bridging the air gap between conductors or causing conductors to come into contact with each other. Examples would include a branch falling over two parallel energized phase conductors built on horizontal cross arms. In this event, a short circuit condition may result if the voltage gradient is sufficient and the branch remains in place for an extended period of time. The primary mechanisms for vegetation-caused outages can be summarized in two categories: - Mechanical tear-down of electric lines and/or apparatus, causing outages - Electrical short circuits or arcs causing overcurrent faults, most often resulting in operation of system protection devices to clear the fault, thereby causing an outage. Frequently, a combination of both tear-down and short circuit mechanisms occur. In fact, either mechanism can occur first, leading to the second. For example, lines torn down by trees (mechanical tear-down) can cause conductors to arc when they hit ground, creating an electrical short circuit. Phaseto-phase tree limb faults (electrical short-circuits) can burn conductors in two, which fall to the ground and sometimes remain energized. 16

The overall contribution of tree or vegetation related outages as a function of all outage causes on the distribution feeder is important to note. Over all utilities, approximately 20% of faults on distribution systems are caused by tree contact. [2] A 1984 Electric Power Research Institute (EPRI) study showed a high correlation between these tree contact faults and adverse weather, a subject addressed hereafter. [3] Significant statistical work and data mining by Duke Power shows similar results as shown in Figure 1. Figure 1: Distribution Faults in the Duke Power System Source: Duke Power, L.S. Taylor, 1995 Data between 1987 and 1990 Duke Power also determined that over 70% of tree outages were caused when an entire tree fell over a line creating a tear-down situation. [4] It is also important to note that 86% of these tear-down conditions were from trees that were outside the utility right-of-way, a figure confirmed by EPRI studies. [5, 6] It also reports that dead limbs or trees cause 45% of tree outages as opposed to outages caused by live trees. The EPRI report is based on data in 1995 1996 for the US and Canada related to tree caused outages covering 17 utilities. This data confirmed that at least 70 to 80% of tree related outages in any given year were due to a tear-down condition, often due to entire tree failure. In that same study, Niagara Mohawk stated that a high percentage of tree related outages or interruptions were 17

caused by uprooted trees. They report that over 80% of permanent tree related faults were caused by trees outside of the right-of-way. [7] 3.3 Mechanical Tear-Down Mechanical tear-down refers to physical destruction or damage of lines, poles, and apparatus without respect to any electrical event. Based on the data above, we conclude that mechanical tear-down represents the primary cause of vegetation-related outages on electric distribution feeders. For this investigation, a ratio of 80% tear-down to 20% other causes will be used for all vegetation-caused outages. Examples of mechanical tear-down causing outages on electric distribution systems are easily found. A frequent cause of outages is large branches breaking and falling over one or more phases of a distribution feeder, often severing the lines and creating a downed conductor. A similar situation occurs when an entire tree falls (referred to as hazard or danger trees), tearing down distribution conductors. In these mechanical tear-down scenarios, an electrical fault may or may not occur, based on local circumstances and system configuration. The physical and weather circumstances at the time of the incident frequently dictate whether an electrical fault accompanies the tear-down or only mechanical damage occurs. In some situations, electric utilities will deenergize distribution lines in advance of major weather events such as floods or hurricanes to prevent electrical damage to apparatus. In this case, because the conductors are not energized, no electrical fault will occur when the mechanical tear-down drops conductors to the ground. However, under most circumstances, a mechanical tear-down condition that results in energized downed conductors also results in electrical faults for at least the first span of conductors to touch the ground on each feeder. If multiple tear-down conditions occur, the entire feeder or a section of the feeder beyond the initial tear-down is deenergized, and subsequent teardown conditions downstream from the open substation breaker, feeder recloser or fuse will not have an associated electrical fault. Repair of a feeder section that has been torn down fundamentally requires the same effort and time whether an electrical fault has occurred or not. Factors contributing to mechanical tear-down include ice accumulation on trees and/or conductors, snow accumulation, and wind generated mechanical forces. There is a high correlation between weather and tear-down events. The best and most recent examples in Texas of mechanical tear-down due to environmental forces are the outages that occurred across many utilities during Hurricane Ike. Hurricane Ike represented the largest outage of CenterPoint customers in history resulting from a single weather event. [8-10] CenterPoint did not, in general, lose energy supply, nor were most substation based apparatus and transformer facilities damaged by the hurricane. For the CenterPoint system in the period following the initial wave of outages, there was an excess of generation available which caused significant operating problems. It was simply not possible to deliver this energy to loads due to excessive outages of 18

components on the distribution system, primarily due to fallen trees and limbs resulting in mechanical tear-down of feeder conductors. The length of outage for the average CenterPoint customer was often dictated by the level of vegetation-induced tear-down at multiple locations on a given electric distribution feeder. In some cases, a single distribution feeder had scores of line conductor spans on the ground and/or poles destroyed as a result of trees falling over lines. For most of these tear-down conditions, no electrical fault occurred because the first fault on the feeder resulted in the operation of the substation circuit breaker for that feeder, thereby deenergizing the entire circuit. Consequently, there was no failure of electrical systems or burn-down of lines but multiple mechanical tear-down conditions that each required physical repair. Remediation efforts were significantly hampered by the need for substantial work by vegetation crews to cut and remove vegetation before access could be gained by electrical workers to repair electric lines and poles. As shown during Hurricane Ike and similar events, the impact of on and off right-of-way hazard and danger trees and their effect on distribution system reliability cannot be overstated. Trees falling during major weather events such as a hurricane, wind storm, or ice storm cause a substantial number of the vegetation related outages in Texas. The Public Utility Commission of Texas (PUCT) separately commissioned reports on hazard and danger trees as well as storm hardening of electric distribution systems related to vegetation outages; therefore, danger/hazard trees will not be addressed in detail in this work. [11] This issue is considered, however, as one important factor affecting vegetation management practices and overall reliability. 3.4 Electrical Causes of Faults A fault condition occurs on an electric conductor when a short circuit condition causes unintended current to flow outside the conductor. For example, a tree limb falling across two horizontal phases of a three-phase distribution line represents a fault condition. Current may flow through the branch causing an arc or explosive event. Electrical faults are generally classified as single-phase or multi-phase short circuit conditions. Most of the exposure miles of typical distribution feeders in Texas are single-phase lines which carry only one energized conductor and a neutral conductor that represents system ground. For these lines, an electrical fault must be a line-to-neutral (or ground) fault, as no other phase conductors are present. Most distribution feeders in Texas are 15 kv-class distribution feeders, with voltages ranging from 11,000 to 14,500 volts measured on a phase-to-phase basis. The most typical distribution feeder in Texas will have a nominal 7,200 volts-to-ground, measured line-to-neutral (~12,500 volts measured phase-to-phase). At 7,200 volts-to-ground, and for conventional construction distances between energized conductors and neutral conductors, a vegetation-caused fault due to trees growing into lines is exceptionally uncommon. In general, a living tree branch bridging the line and neutral conductors will not draw sufficient current or act to create an arcing fault condition. Momentary operations and/or lockout of 19

system protection devices are uncommon for this configuration. While clipping or electrical pruning of new growth vegetation near distribution lines occurs due to the microamp to milliamp currents that flow on the vegetation, these events do not often result in an arcing fault, do not result in substantial current flow, and do not represent an interruption of service to customers since no line protection devices will operate to deenergize any part of the feeder. Experimental work has shown that 7,200 volts-to-ground from a phase conductor to vegetation represents an insufficient voltage gradient to create a fault condition through vegetation over any significant distance. [12] Experimentation by Goodfellow summarized in Figure 2 shows the probability that a fault will occur based on voltage gradient across a tree branch. For low voltage gradients (less than 2 kv per foot), an electrical fault is extremely unlikely. As the voltage gradient across a branch increases, the probability of an electrical fault occurring also increases. If a branch spans a 7,200 volt feeder with a phase spacing of six feet, the voltage gradient across the branch will be 1.2 kv. As Figure 2 shows, the probability of burning of a limb with a subsequent high current arc at this voltage gradient is very low. At voltage gradients below 2 kv per foot, the probability of a fault occurring is almost zero. This is because low voltage gradients do not cause limbs to burn or arc, which must happen before current can flow. The charring or burning of a limb takes time. Figure 3 shows that the time required to create a fault condition at a voltage gradient of less than 2 kv per foot is very high. This means that if a tree only contacts a single-phase energized conductor or bridges the energized conductor and neutral conductor, absent other compounding circumstances, a fault condition will not occur. A detailed analysis of the vegetation-related electrical mechanisms causing distribution faults is found in Appendix A. Figure 2: Fault Causation due to Voltage Gradient Source: Goodfellow, 2000 20

Figure 3: Time Required to Create Vegetation Faults Source: Goodfellow, 2000 There is a common misunderstanding that when trees grow into power lines, surrounding the lines below and above, there is constant contact between the energized conductors and the trees at many thousands of points. I have documented a different conclusion based on the experience of utility linemen and the observations of researchers. The picture shown in Figure 4 is typical of the tunnel created around energized conductors where trees have grown around the conductors to envelop the line. The mechanism can be described in lay terms as follows. Figure 4: Typical "tunnel" effect from an energized conductor passing through vegetation Source: Picture by B. Don Russell, 2010 21

As new vegetation growth touches an energized conductor, micro-amp currents begin to flow which heat, destroy, or clip the vulnerable growth. Moisture in new growth is heated, and growth ends are destroyed. This does not represent a burning of limbs and, in general, no fire danger exists from this specific phenomenon even during drought conditions. As a consequence, conductor contact with this new growth and small limbs is much reduced as the conductor carries out a self-trimming process. The subsequent growth of the tree will therefore be to the sides and then will grow back over the conductors leaving the characteristic tunnel shown in Figure 4. Other mechanisms at play in creating this phenomenon are the mechanical forces that occur around the conductor due to conductor movement and due to the movement of branches. These forces can be considerable during wind storms. The movement creates mechanical trimming of certain branches, resulting in the characteristic tunnel effect. As a result of the described mechanisms, the significant majority of exposure miles of distribution conductor has no contact with vegetation, even though full tree growth under and around and over the conductors on the right-of-way may exist. Once this characteristic tunnel has been created and branches have overgrown and become dense above the conductors, little sunlight reaches the tunnel area, and regrowth of the inside branches is inhibited. Again, contact with conductors inside the tunnel is mitigated, and for much of the feeder, often eliminated. Direct tree contact with single-phase 15 kv class conductors is not a statistically significant cause of faults. The experience of distribution engineers who study the number and type of faults documented on feeders confirms this finding. [13] If each instance of new growth of trees into distribution conductors actually caused an electrical arc and fault condition, most distribution feeders in operation in overgrown vegetation areas would be faulted on a virtually continual basis. This is, however, not the case. Experimentation and direct measurement of feeders with heavy vegetation growth has been performed by the Power System Automation Laboratory at Texas A&M University. This has shown that for feeders in heavy vegetation with many exposure miles of conductor, it is very rare for the feeder to experience an arcing fault and that momentary, self-healing interruptions due to tree contact with single-phase conductors are also rare. [1] Phase-to-phase vegetation contact on a distribution system represents a substantially higher voltage gradient and if a contact is sustained, frequently does result in the operation of system protection devices, leading to an outage condition. Normal spacing between phases (e.g. four feet) on a 12.5 kv phase voltage would result in greater than 3 kv per foot voltage gradient. As shown in Figures 2 and 3, the probability of a fault occurring is more significant as phase voltage increases. This becomes an important issue as the trend toward higher distribution voltages increases to support increasing load levels. If a tree limb is weighted by ice and bridges two phases of a feeder, or if it breaks and falls from a tree falling over two phases in a horizontal constructed distribution feeder, an arcing fault with significant follow-on current may occur after several minutes. An infrequent, but well documented, occurrence is a branch falling over two phases and either burning clear or, alternatively, causing a permanent fault. If 22

the branch is not removed from the line either through damage to the branch (e.g. burning) or other mechanical action, the resulting high current event will likely cause a lockout of a recloser or substation breaker, leading to an outage. An actual example of this mechanism is described in the case studies of Appendix B. It is commonly thought that electrical current conducts uniformly through the cross section of a tree branch, causing a rapid short circuit event and resulting in an outage. There is no scientific support for this mechanism. When a broken branch or other vegetation bridges line-to-neutral or phase-to-phase conductors, various electrical activity is initiated, including carbonization of the surface of the vegetation with a progression that may, after substantial time, lead to a low impedance path and possible arcing fault. Again, this mechanism is described in detail in Appendix A and in reference [14]. In general, a voltage gradient of greater than 2 kv per foot is needed to generate an arcing fault. Additionally, the tree branch must contain enough moisture to allow an initial high-impedance conductive path to form. [1] Another mechanism for electrical fault occurrence involves branches falling and pushing one phase conductor into another phase conductor or system neutral, thereby causing a fault. Similarly, weighting caused by ice or snow pushing vegetation down onto conductors causes faults. In these cases, the construction architecture of the distribution conductors (e.g. vertical or horizontal) dictates the nature of the fault. In some situations, an energized line over a neutral several feet below can be pulled down, resulting in metal to metal contact between the line and neutral conductors, thereby generating a fault. Similarly, a phase-to-phase fault may occur for either vertical construction or horizontal construction where tree branches push or pull phases together or bridge phases creating a conductive path. In summary, vegetation-caused electrical faults frequently occur when the natural insulation properties of air are lost after vegetation reduces the effective resistance between the energized conductors and/or ground. Another different but similar mechanism is the metal to metal contact that occurs if phase conductors and/or neutral conductors are pushed together by vegetation. It should be noted that many, if not most, electrical short circuit conditions caused by vegetation do not cause a reported outage but may result in a momentary interruption. However, utilities generally do not capture and document the frequency or magnitude of momentary interruptions from any cause, and are unable to distinguish vegetation-related momentary interruptions from those caused by other factors. [14] It is important to note that momentary interruptions, by definition, result in a loss of power to all affected customers, which may last up to several minutes. [15] A detailed explanation of the behavior of electrical faults from vegetation is included as Appendix A. 23

3.5 Fire and Electrical Injury Hazards Several hazards occur when power lines fail and fall to the ground due to either mechanical tear-down and/or electrical faults. As described in sections 3.3 and 3.4, when a tree falls and rips conductors from insulator supports and falls to the ground it often remains energized. A 15 kv class feeder can have downed energized conductors remain on the ground in an energized and arcing condition for many minutes until a sustained over-current fault causes operation of protective devices. Arcing from these lines can cause fires. Similarly, a limb across two phases in the air can cause intermittent arcing and burning of the limb. Falling pieces of the limb and burning embers may result. An example of this event is included in Appendix B. Both arcing downed power lines and burning limbs represent competent ignition sources for fire if ground conditions are susceptible. Sufficient fuel load must exist and moisture conditions must be suitably low for combustion to occur. However, sustained low rainfall levels often result in drought conditions that create an extreme fire hazard. These conditions have been experienced in the last two years in Texas and have created a fire hazard that continues today. In addition to concerns over fire causation, downed power lines that remain energized represent an electrical hazard to the public. In many cases, protective devices, cannot see downed power lines which may not draw sufficient current to operate breakers, reclosers, and fuses. Fires may occur and electrical hazards may persist beyond the control of the utility. However, when possible, attention must be given to mitigate and eliminate the conditions that may cause downed power lines. 3.6 Learning Points 1. Trees and other vegetation represent less than 20% of all fault causation for non-storm conditions. 2. Mechanical tear-down is the primary (e.g. 80%) cause of vegetation outages. This is exacerbated during storms and/or high winds which cause trees to fall. 3. Electrical contact between a single conductor and live branches is rarely the root cause of a vegetation-caused outage. 4. Single-phase vegetation faults for 15 kv class or lower distribution voltages are rare due to the relatively low voltage gradient from line to ground. 5. Arcing vegetation faults on 15 kv class single-phase feeders are rare absent mechanical forces causing direct phase to neutral (metal to metal) contact. 24

6. Higher voltage distribution feeders (e.g. 25 kv, 35 kv) have an increased probability of electrical faults due to vegetation because of the higher voltage gradient. 7. Phase-to-phase vegetation faults occur on 15 kv feeders if two conditions are met. (a) The vegetation (e.g. branch) must bridge phases in a mechanically stable way over a sufficient time period to create an arc path by charring and burning the branch (generally requires solid contact on the order of minutes). (b) The vegetation must not burn or fall free before a permanent outage occurs (e.g. arcing fault initiating protective device operation). 8. Downed energized electrical conductors represent a fire hazard and an electrical hazard to the public. 25

4.0 Case Studies - Experience of Utilities Documenting Vegetation Management Effectiveness Controlled experiments conducted by electric utilities to evaluate vegetation management practices are rare. This is due to a significant number of factors including the long duration over which these experiments must be conducted and the liability associated with allowing vegetation to continue in contact with lines long after the incidence of outages has markedly increased. Consequently, documented examples of how particular vegetation management practices have directly affected the performance and reliability of distribution feeders of a given utility are most important. Two examples have been found which provide us with great insight and understanding of the effects of specific vegetation practices on the reduction of the frequency and duration of outages. 4.1 Example One Seattle City Light The first example comes from the experience of Mr. Bernie Zeimianek who is the director of Energy Delivery Operations for Seattle City Light (SCL). Mr. Zeimianek was interviewed for the purposes of this report. Based on discussions with Mr. Zeimianek and based on information he has published in Transmission and Distribution World magazine, we have summarized the SCL experience. [16] For multiple reasons, Seattle City Light had a period approaching almost a decade where regular vegetation management was not practiced. During this time, very few funds were expended on vegetation management. Little work was done beyond trouble men and linemen clearing vegetation in the immediate area where an outage or event had already occurred. One can describe the process used during this period by SCL as purely reactive, clearing vegetation after a fault had occurred but with no proactive approach or periodic trimming of trees and vegetation. Mr. Zeimianek brought his considerable industry experience to SCL and, with broad support, initiated a comprehensive vegetation management program in 2006. The hallmark of this program was a four year trim cycle over the entire SCL system. Mr. Zeimianek calls this a system wide haircut. SCL operates a 26 kv system, which creates the potential for higher voltage gradients than 15 kv class systems. As a result, Mr. Zeimianek chose to fully trim utility easements to a 10 to 15 foot clearance from the line infrastructure. This effectively removed all intruding vegetation, right-of-way trees, etc. The data benchmark for unreliability cited by Mr. Zeimianek was the following. After a decade of no trimming, except during repairs of outages, SCL experienced 12,000 vegetation-related outages per year. In the context of the size of SCL, this represents a very high incidence of vegetation-related outages. Mr. Zeimianek reports that many outages during this period could be referred to as burners, where vegetation-related electrical activity or arcing vegetation caused outages and/or damage. He also noted that there was a very strong correlation between the number of calls and outages and wind levels greater than 20 MPH. The dramatic improvement of SCL after initiating a system wide haircut and a four year trim cycle cannot be overemphasized. After four years of trimming, Mr. Zeimianek reports that the number of tree 26