000000000001012442
000000000001012442
000000000001012442
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.
ELECTRIC POWER RESEARCH INSTITUTE 3420 Hillview Avenue, Palo Alto, California 94304-1338 PO Box 10412, Palo Alto, California 94303-0813 USA 800.313.3774 650.855.2121 [email protected] www.epri.com
This is an EPRI Technical Update report. A Technical Update report is intended as an informal report of continuing research, a meeting, or a topical study. It is not a final EPRI technical report.
NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or e-mail [email protected]. 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 document was prepared by Coplanar Consulting, LLC Yardley, PA. Principal Investigator M. Lebow This document describes research sponsored by the Electric Power Research Institute (EPRI) This publication is a corporate document that should be cited in the literature in the following manner: Performance-Focused Maintenance for Distribution Substations: Survey and Guide with KPIs and Algorithms for Living and Predictive Maintenance. EPRI, Palo Alto, CA: 2006. 1012442.
iii
ABSTRACT
Performance Focused Maintenance (PFM) was introduced in 2005 as a generic concept to help maintenance managers cope with the competing objectives of maintaining service standards while containing costs. This report further develops the principles of PFM and integrates them into the larger power delivery asset management process. As the name implies, PFM focuses on performance and this report develops and demonstrates the steps required to establish proper performance measures and KPIs for PFM implementation. The concepts of PFM are illustrated through an application to distribution substation maintenance. Issues and gaps in the existing level of knowledge are identified. The report concludes by recommending several efforts required to bring PFM to an application-ready state.
ACKNOWLEDGEMENTS
EPRI wishes to acknowledge the support and assistance of Mr. Bhavin Desai.
vii
CONTENTS
1 INTRODUCTION ....................................................................................................................1-1 Background ..........................................................................................................................1-1 Drivers for PFM ....................................................................................................................1-2 Costs and Service Quality Balance................................................................................1-2 Power Delivery Asset Management ...............................................................................1-3 Aging Asset Base...........................................................................................................1-4 Resource Limitations......................................................................................................1-4 Project Scope.......................................................................................................................1-5 Report Organization .............................................................................................................1-6 Chapter References .............................................................................................................1-6 2 INTRODUCTION TO POWER DELIVERY ASSET MANAGEMENT.....................................2-1 What is an Asset? ................................................................................................................2-3 What is Asset Management? ...............................................................................................2-3 Formal Definitions ..........................................................................................................2-4 Power Delivery Asset Management .....................................................................................2-5 Asset Management Premise ................................................................................................2-6 Implications of an Asset Management Approach.................................................................2-6 Benefits of Asset Management ............................................................................................2-7 Requirements for Asset Management..................................................................................2-9 A Power Delivery Asset Management Diagram ...................................................................2-9 Goals and Policies .......................................................................................................2-11 Performance Assessment ............................................................................................2-12 Summary of Asset Management Concepts........................................................................2-13 Chapter References ...........................................................................................................2-13 3 PERFORMANCE FOCUSED MAINTENANCE AND POWER DELIVERY ASSET MANAGEMENT.........................................................................................................................3-1 The Core Maintenance Concept ..........................................................................................3-1 The Maintenance Process ...................................................................................................3-1 Maintenance Process Diagram ............................................................................................3-3 PFM Characteristics.............................................................................................................3-6 Performance Measurement..................................................................................................3-7 4 PERFORMANCE MEASURES ..............................................................................................4-1 Background Definitions ........................................................................................................4-1 Performance Metrics ............................................................................................................4-2 Guidelines for Developing a PFM Measure .........................................................................4-3 Key Performance Indicators.................................................................................................4-3 Distribution Substation Performance Measures...................................................................4-4
ix
PFM and Performance Measures ........................................................................................4-8 Secondary KPIs...................................................................................................................4-9 Chapter References ...........................................................................................................4-12 5 PERFORMANCE FOCUSED MAINTENANCE FOR DISTRIBUTION SUBSTATIONS........5-1 Distribution Substation Maintenance....................................................................................5-1 PDAM and Distribution Maintenance ...................................................................................5-3 PFM Fleet Management.......................................................................................................5-3 Linking Maintenance to Equipment Reliability......................................................................5-5 Fleet Management .............................................................................................................5-12 Power Transformer Fleet Management .......................................................................5-13 Equipment Performance Data Base...................................................................................5-16 Chapter References ...........................................................................................................5-18 6 CONCLUSIONS AND RECOMMENDATIONS FOR ADDITIONAL WORK ..........................6-1 Conclusions..........................................................................................................................6-1 Recommendations ...............................................................................................................6-1 Data................................................................................................................................6-1 Equipment Models .........................................................................................................6-2 Pilot Studies ...................................................................................................................6-2
1
INTRODUCTION
Background Previous EPRI work (Technical Update: Performance Focused Maintenance) 1 introduced Performance Focused Maintenance (PFM) as a new approach to maintenance intended to meet the continuing competitive challenge of improving maintenance performance while reducing costs. It was proposed that PFM could provide a means to go beyond Reliability Centered Maintenance (RCM) and Condition Based Maintenance (CBM). The stated goal of PFM was to strengthen and build on rather than replace a utilitys current maintenance practices. This previous effort developed PFM at a generic, conceptual level. The objective of the work reported upon here is to further develop the PFM concepts and integrate them into the larger set of asset management principles that are guiding power delivery organizations today. Conceptually, PFM is not one particular formula or technique but rather a maintenance philosophy that merges a conventional maintenance approach with asset management. It can complement existing programs, including RCM and CBM, but does not rely on them. Rather PFM is an approach to establishing the contribution of maintenance to higher-level goals.
2 RCM analysis can be time-consuming and many utilities have been reluctant to undertake 3 system-wide application. Similarly, CBM (see Substation Life Extension Guidelines for complete definitions), although recognized as one of the best maintenance approaches, relies on either on-line monitoring or extensive data collection and analysis. PFM provides a framework that can accommodate both approaches but also provides a methodology to improve maintenance maximizing the benefits from existing resources and data. Figure 1-1 shows that PFM considers not only equipment functionality and reliability but also the relationship the equipment has to the organizations performance goals and objectives. This approach allows for a broader view of the contributions of the maintenance process and focuses resources where they best support higher level objectives.
Traditionally, maintenance has been based on preserving the physical assets and striving to achieve the highest possible equipment reliability. This often emphasizes preventive maintenance schedules that contain generic tasks and frequencies and which may lead to ineffective maintenance either through inadequate or unnecessary work. PFM would suggest achieving a level of maintenance that matches the strategic service level requirements by understanding the total performance contributions of the equipment and the maintenance program. The goal of PFM is to help maintenance and asset managers direct their limited resources to the maintenance tasks that best contribute to reaching the organizations business goals.
1-1
Three objectives were established and achieved in this second phase of the evolution of PFM. Merge the concepts of PFM with the guiding principles and precepts of power delivery asset management. Expand on the definitions and measures of performance in relationship to PFM. Illustrate the use of PFM concepts through application to distribution substation maintenance.
Finally, this work proposes steps to continue development and enhancement of the PFM concepts.
Figure 1-1 PFM Not Only Considers Maintaining Equipment but Also the Equipments Contribution to Performance Goals and Objectives
Drivers for PFM Costs and Service Quality Balance The central area of conflict for power delivery organizations today is defined by the opposing objectives of cost containment and preservation, or in some cases improvement, of the quality of
1-2
service. This struggle is of particular importance to those operating distribution networks, since distribution networks have a significant influence on both the quality and the costs of the power supply. Distribution substations, as defined for this project, are often not the major focus of distribution system maintenance attention. New construction and the maintenance of lines and equipment outside the fence usually consume the larger portion of the budget. The lower initial cost of distribution substation equipment as compared to transmission equipment further reduces the visibility of its maintenance. Consequently, distribution substations maintenance budgets often are assigned a lower priority and become one of the first candidates when budget reductions are considered. The high costs of performing an RCM analysis or collecting data for CBM makes these approaches less attractive for distribution substations. PFM offers a way to demonstrate the results of distribution substations maintenance and improve its effectiveness. Power Delivery Asset Management Power delivery equipment owners and operators have become aware of some of the limitations of conventional business approaches. Maintenance operations often are disproportionately affected by these shortcomings. Depending on circumstance and degree, these limitations often include preconceived budgets, emphasis on costs rather than stakeholder values, focus on shorter-term equipment issues rather than long-term global issues, inconsistent business cases, difficulty in compiling and up-dating asset data, inadequate horizons for long-term planning, reliance on projection of historical results for future performance, and insufficient attention to risk and vulnerabilities throughout the remaining asset operating life. Formulating and presenting maintenance programs in a power delivery asset management (PDAM) framework can help address these limitations. The principles of asset management4 will be discussed in detail in the following chapter but some explanation at this point will be helpful in introducing how PFM fits into PDAM. In essence, asset management is the establishment and execution of a series of interrelated processes that assure that all decisions related to the allocation of resources are evaluated against, aligned with and made to optimize the achievement of the organizations goals for financial and operational performance and tolerance for risk exposure. Some key characteristics of PDAM are: Asset management is driven by policy, goals and objectives based upon performance. Asset management entails the translation of policies and plans into optimized investment strategies and the translation of investment strategies into optimized program delivery. Investment choices and decisions on allocating and applying resources are policy and performance-driven. Investment choices and decisions on allocating resources are based upon explicit tradeoffs among projects, programs or strategies. Asset management takes a long-term view of infrastructure performance and cost. Asset management is proactive. For asset management to be most effective, a process must be in place to guide longer-term maintenance direction. For maintenance to be effective there must be processes to ensure the direction over time tracks an organizations mission and goals. PFM is the process that
1-3
formulates the maintenance programs to support the goals and objectives that are the basis for power delivery asset management and asset management assures their execution. Of course, maintenance currently is a well-established process in all power delivery organizations. In fact, some confuse good maintenance practices with asset management. Maintenance is an integral part of asset management, but only a part. In a PDAM approach, resource allocation decisions are tightly linked to policy objectives, and based on a quantitative understanding of the results obtained for every dollar invested. This includes maintenance decisions and PFM attempts to address the question: How does maintenance affect asset performance in relation to company performance? Aging Asset Base Many in the industry believe that a significant percentage of the power delivery equipment base is at or close to the end of its useful life, the so-called design life. There is no formal definition of design life but the general usage is that it is the age beyond which risk of failure will become increasingly unacceptable. The concept arises from the fact that the original equipment designers and purchasers did not expect the equipment to be in service much beyond that age. Equipment engineers commonly use an age of forty years to estimate design life for power delivery equipment but there is no technical basis for this number and there are many examples of equipment functioning reliably well beyond that age (and, of course, many that never achieved that age). However, it is generally accepted that the risk of equipment condition deterioration and wear-out failure increases as equipment is used and ages and approaches some end of its useful service life. Many of the power delivery systems in the United States experienced a rapid expansion in the 1960s and 1970s corresponding with significant national economic growth and increased electric consumption. Much of the equipment installed in that timeframe is still in service and the equipment installed during that peak expansion is now 30 to 40 years old. Much of the recent literature concerning aging transmission assets deals with power transformers because they are usually the single most expensive component in the delivery system but the same situation exists for distribution system equipment. Replacing this significant population of older equipment will require a large capital investment and hence the interest in the aging asset problem. This is an issue for consideration in any maintenance planning process since maintenance can be expected to influence equipment service life and failure rates. Resource Limitations In the last decade there has been a net decrease in the utility labor force. In step with the power delivery industrys rapid expansion of the equipment base in the 1960s and 1970s, a similar increase occurred in the size of the workforce. More recently, the economic pressures resulting from deregulation and a series of mergers have resulted in a significant reduction in the number of people employed in the utility industry, both skilled craftsmen and engineers. It has been estimated that over half of all electric utility workers are 45 or older5. This reduction, coupled with the approaching retirement of those remaining, means that skilled technical, craft and engineering expertise is in increasingly short supply. This shortage exacerbates the challenges of dealing with an aging asset base and the ever more complex issues of operating and maintaining systems under pressure to contain costs and maintain quality
1-4
service. Consequently, most utility equipment experts are fully engaged in dealing with emergent problems and system growth and have little time to assess equipment fleets and improve maintenance programs. Similarly, there are resource limitations on gathering the detailed historical and condition data and on-site inspection and testing results one would ideally wish to review to assess an equipment population. Project Scope PFM recognizes that, at the highest levels, maintenance processes should be very similar for most utilities. However, PFM also recognizes that the approach to and application of these processes will differ from company to company due to individual circumstances, including the wide range of customer requirements, electric infrastructures and maintenance organizations. The initial work presented here developed a generic approach suitable for the distribution business. KPIs for performance and predictive based maintenance for distribution substations have been defined. Based on assessment of industry needs, subsequent efforts will customize the methodology for specific equipment types and applications. The goal of the project is to provide a maintenance framework that integrates many of the technical, economic and managerial concepts that have been a foundation of maintenance in the past as well as recently introduced concepts and ideas. The results will incorporate the concepts of modern asset management and integrate many previous EPRI activities, assisting utilities in building custom maintenance strategies that meet their business needs. PFM has been presented as a broad approach suitable across the power delivery industry. The first part of this report will continue development of this broad PFM concept. Later, focus will be directed to distribution substations. It is recognized that there is no industry standard for the designation Distribution Substation and utilities have different boundaries as to what constitutes their distribution and transmission system. The upper boundary varies considerably while the lower boundary is usually the customer meter. Therefore, while some utilities will consider high voltage equipment as part of the distribution system and others will not, every utility has lower voltage equipment in its distribution system. It is this lower voltage substation equipment, generally considered by all to be distribution that is the focus of this project. In most technical discussions, distribution equipment voltage classes are usually considered to be below 50 kV although transformers with greater high side voltages are often included in distribution substation maintenance programs. From one perspective, there are many similarities between transmission and distribution substations. The major equipment types are the same: transformers and breakers. The maintenance objectives are also the same: maintain the desired performance levels as economically as possible, but there are some real differences. Distribution equipment, as defined above, is much less expensive. The lower replacement costs make it hard to justify the level of maintenance, monitoring or testing that transmission (higher voltage and larger MVA) equipment warrants. Distribution stations are configured differently and are, obviously closer to the customer. Problems in distribution stations have a greater chance of affecting customer reliability indices and switching options are often more limited. Most distribution stations are unmanned and often remote. They often do not get as much attention as transmission stations. There are differences in the basic equipment designs also; metal clad switchgear and unit substation transformers for instance. Because of their smaller sizes, commodity marketplaces
1-5
and lower prices, design margins can be less in distribution equipment. These, and other reasons, mean that a truly performance focused maintenance approach would be different for both classes of station. After all, the drivers and constraints are different. In addition, the fact that there are such large numbers of very similar equipment in distribution may provide a real possibility of developing statistically meaningful hazard rate and life expectancy curves based on real-world data. Report Organization This report is divided into six chapters. In addition to this introductory chapter, included are: Introduction to Power Delivery Asset Management Performance Focused Maintenance and Power Delivery Asset Management Performance Measures Performance Focused Maintenance for Distribution Substation Conclusions and Recommendations for Additional Work
Chapter References 1. Technical Update: Performance Focused Maintenance. EPRI, Palo Alto, CA. 2005. 1008673 2. Reliability Centered Maintenance (RCM) Technical Reference for Substations. EPRI, Palo Alto, CA. May 1996. Report TR-106418. 3. Guideline for the Life Extension of Substations. EPRI, Palo Alto, CA. 2000. 1000031 4. Guidelines for Power Delivery Asset Management. EPRI, Palo Alto, CA: 2005. 1010728 5. Ray, Dennis and Bill Snyder. Strategies to Address the Problems of Exiting Expertise in the Electric Power Industry. IEEE International Conference on System Sciences, 2005.
1-6
2
INTRODUCTION TO POWER DELIVERY ASSET MANAGEMENT
Maintenance is an integral part of asset management. A best practice Power Delivery Asset Management (PDAM) implementation incorporates a well-constructed maintenance process implementation. A well-constructed maintenance process is built on and adheres to the requirements of asset management. Together, they guarantee integrated planning, assessment, budgeting, execution and evaluation of asset decisions across the assets life cycle. This chapter outlines the basics of power delivery asset management and provides a context for a detailed discussion of PFM. For the foreseeable future, utilities will need to manage an array of potentially conflicting business objectives, including the need to maintain competitive economic performance, improve customer satisfaction, maintain high reliability, address regulatory uncertainty, and comply with increased environmental regulation. The result is that many utilities are considering or have moved towards implementing formal asset management concepts and driving decision-making based on minimizing equipment life-cycle cost and maximizing benefits. A structured asset management approach has been successful in many other industries and, when properly adapted to utility needs, can provide the processes and tools to develop the most effective programs for building, operating and maintaining todays power delivery infrastructure. Power Delivery Asset Management begins with the fundamental premise that all asset management decisions made by utilities should contribute to stakeholder values, as set forth in the organizations goals and policies. PDAM applies this premise in decision processes at every level of the organization. The resulting alignment of decisions with criteria and value measures derived from the asset owners or senior managements direction ensures that every asset management and resource allocation decision consistently supports the organizations strategic objectives and delivers value to the stakeholders. This chapter describes the principles and fundamental processes of Power Delivery Asset Management in order to provide the background and perspective for understanding how PDAM and PFM complement each other. Much of the material in this chapter is derived from 1 Guidelines for Power Delivery Asset Management and those interested in more detail are referred to that document. The concept of asset management has been fundamental to the business of electric utilities throughout their history. Companies have always endeavored to manage their assets, employees, capital, and equipment to deliver as much perceived value as possible, and these efforts have been highly successful. However, several aspects of the traditional ways of conducting business in the electric power industry have changed. Many organizations have unbundled traditional vertically integrated utility functions via the sale of assets or entire operations, or by redefining roles and responsibilities (see Figure 2-1). In some circumstances, the roles are assigned to different enterprises, while in other cases, organizations within the same enterprise now perform these functions. Formal service level agreements have been established to define the roles and
2-1
obligations of the three parties. Even where no separation has occurred, recognizing these distinct roles is helpful when exploring asset management concepts.
Service Provider
Strategy Planning Risk Management
Each of the distinct functions in Figure 2-1 has a well-defined role. The asset owner is the regulatory license-holder represented by the highest levels of management within the organization that owns or, in the case of governmental agencies, directly controls the assets. The asset owner may or may not be a part of the organization that operates and maintains the assets. Owners directly interact with key stakeholders (e.g. customers, shareholders, regulators, employees and financial agencies) and asset managers. The asset owner sets the business goals and policies, parameters of risk, cost and performance, and the budget for the organization. For example, the asset owner sets annual capital and operating budgets, customer satisfaction measures, and fuel mix risks. When asset management practices are applied in an organization below the enterprise level an operating unit for example, the senior management to which the organization reports carries out the role of asset owner. (For convenience, this report applies the term asset owner to both cases.) The asset owner determines the operating context for the asset manager, focusing on corporate governance and goals, regulatory issues, and other stakeholder relationships. The asset manager develops asset strategy and policy and directs risk management, investment and maintenance planning (not work scheduling), and contract management. The asset manager sets the policies and procedures for the service provider(s) and decides how and where money is to be spent for both capital improvements and maintenance. For example, the asset manager sets feeder outage goals, equipment maintenance intervals, and replacement criteria. In short, the asset manager decides what to do and in which budget cycle to do it.
2-2
The service providers are then able to focus on core skills of scheduling personnel to deliver programs efficiently and effectively to meet defined service levels. They provide and schedule resources to perform work on the assets. For example, service providers set maintenance staffing levels, tool requirements, and work schedules. Rather than decide where or how to invest budgets, the service provider decides how to do work. As utilities develop these new business models and the technology to support their engineering expertise to meet the emerging challenges, there has been a gradual change in focus towards power delivery asset management practices. However, asset management is about more than maintenance or capital investment issues that are usually the first areas for attention. At its best, asset management represents the ability to understand and manage the trade-offs between risk, cost, and performance in order to optimize the financial and service performance of the three distinct roles asset owner, asset manager, and service provider that result from a fundamental approach to managing assets.
What is an Asset?
An asset is any resource that is important to an organizations functions and requires management. The organizations assets are used to service and supply end users or to facilitate performing such services. Asset owners acquire, operate, and maintain assets to support service delivery. Therefore, an asset possesses service potential or future economic benefit. In the power delivery industry, physical assets such as transmission and distribution system equipment are the most commonly considered. However, the more comprehensive application of asset management principles might also consider time, people, data and knowledge, and know-how to be assets. Fundamentally, an asset has value that persists, and often changes, over time. This implies that assets have both a useful and an economic life, which may not coincide in length. For practical purposes, only assets with significant value are considered in the asset management process. The management of financial assets is not within the scope of PDAM.
2-3
Assumes a long-term view of infrastructure performance and cost Is forward looking and seeks to predict and anticipate Is pervasive, affecting the business practices of every organizational element involved in the functions to which it is applied As a process, effective asset management: Has executive support Translates policies and plans into optimized investment strategies, and translates investment strategies into optimized program delivery and procedures Requires policy-driven, performance-driven choices and decisions on allocating and applying resources Requires that investment and decisions on allocating resources are based on explicit tradeoffs between projects, programs, or strategies Requires that risks are fully assessed and managed Develops organizational roles and responsibilities regarding asset management Promotes consistent practices across various organizations within an enterprise Documents plans (for normal operations and responds to unexpected events and changing circumstances) that maintain focus on goals and objectives) Is interdisciplinary, combining both engineering and economic tools and processes so that business functions become an integral element of operations Requires effective communication within and outside the organization, and established mechanisms for performance review and adjustments to correct for deviation from desired results As a set of technical tools, asset management: Requires effective management systems Requires the best available, current, and accurate information on assets and asset performance Requires well-developed decision support analyses for evaluating tradeoffs and prioritizing actions
Formal Definitions Numerous organizations have published formal definitions for the asset management process. Following the recent development and adoption of infrastructure asset management, most definitions originate from overseas organizations concerned with public service oversight. Definitions include the following: Systematic and coordinated activities and practices through which an organization optimally manages its physical assets and their associated performance, risks and expenditures over their lifecycles for the purpose of achieving its organizational strategic plan. The Institute 2 of Asset Management
2-4
Asset management is a systematic process to guide the planning, acquisition, operation and maintenance, renewal and disposal of assets. Its objective is to maximize asset service delivery potential and manage related risks and costs over their entire lives. The 3 Government of Victoria A comprehensive and structured approach to the long-term management of assets as tools for the efficient and effective delivery of community benefits. AUSTROADS4 Other definitions abound but these are particularly interesting because civil infrastructure, for example roads, represents a large investment, distributed over large areas, and are long lived. They have many similar characteristics to the power delivery infrastructure.
2-6
integrated set of business processes and decision support tools, which have interfaces with the asset managers key stakeholders (especially the asset owner and service providers). The asset manager uses these processes and tools to manage asset lifecycles, as well as to manage the key interfaces and workload of the various internal and external service providers. These processes and tools and their elements can be summarized under the groupings listed below. Communications Accurate and timely information flows to both owner and service provider Documented, consistent decision-making Performance measures, standards, and benchmarks Useful outputs, effectively presented Data Collection and Analysis Inventory of all assets Valuation of assets Quantitative condition and performance measures Measures of how well strategic goals are being met Usage information Strategy Evaluation of asset performance and influencing factors Risk goals and measures Evaluation of lifecycle costs and benefits Performance-prediction capabilities Planning Engineering and economic analysis tools Alternative analyses procedures Project prioritization procedures Evaluation to balance short- and long-term objectives Implementation Contract management Results monitoring and reporting Continuous feedback procedures
2-7
and techniques to apply in policy-making, planning, project selection, program tradeoffs, data gathering, and management system application, each of which is aligned with the organizations higher-level goals. The benefits of asset management include the following: Assure that all asset decisions are policy driven Build, maintain, and operate facilities most cost-effectively Achieve desired performance levels Optimize long-term benefit/cost ratios Allocate available resources efficiently to support the organizations overall goals and policies Measure and focus on performance and outcomes Improved repeatability, credibility, and accountability for decisions
Asset management links customer and regulator expectations for system condition, performance, and availability with system management and investment strategies. A complete asset management process reports on progress made in achieving performance measures derived from asset owner goals and also evaluates the business processes used relative to the goals and performance criteria. Furthermore, the impact of alternative actions and investment strategies on the ability to realize expressed goals may be readily determined and evaluated. The focus is on assets (i.e., data, people, and physical resources) and system performance, including return on investment, economic efficiency, accountability, opportunity costs, risk exposure and future requirements. This broad approach to resource allocation and decisions can provide greater benefits to the organization and all stakeholders. Asset management not only aids in the decision-making process, but also provides for a factbased dialogue between asset managers and other stakeholders, government officials, and customers concerned with day-to-day operations. This results from the accessibility of relevant, objective, and credible information to all participants in the decision-making process. As such, decisions can be based on detailed input regarding available resources, current system condition and performance, and estimates of future performance. The information underlying the asset management processes raw data and results generated from analysis ensures an improved understanding of the economic tradeoffs, return on investment, performance impacts, and accountability. Asset management provides ready access to quantitative and qualitative data and allows decision makers to more readily identify and focus on key issues. Furthermore, the ability to weigh and articulate the impact of choosing one alternative over another through what if analyses is enhanced. The documentation explaining the selection of a particular strategy is also improved. A documented, reproducible, systematic approach can enhance communication between stakeholders and provide the asset manager a defensible rationale for capital investments and other actions.
2-8
2-9
Customers Regulators
Asset Owner
Formulate Policies
Asset Manager
Develop Program
Implement Results
Monitor
Service Provider
Figure 2-2 Power Delivery Asset Management Framework
Adding some detail to Figure 3-2 yields the power delivery asset management model shown in Figure 3-3. Its blocks can be readily correlated with the preceding Figure and the boundaries of the three parties responsibilities can be deduced. Performance criteria and assessments are the key to determining future actions. Processes for monitoring how well expectations are met are evident. This model fits most aspects of the power delivery business. The PFM process model
2-10
introduced in the next chapter will have clear parallels to the PDAM model as would be expected from an integrated approach.
Set Goals Policies
Budgets
Stakeholders
Asset Inventory
Performance Criteria
Assess Performance
Evaluate Optimize
Implement
Monitor Results
Goals and Policies The asset owner is the initiator for PDAM. The owner sets the business parameters, risk boundaries, and operating context for its assets for the operational and longer-term horizons. The asset owner also sets the operating context for the asset manager and, for power delivery, focuses on corporate governance as the regulatory license-holder. In general, the owner is represented by the highest levels of management within the organization that owns or, in the case of governmental agencies, directly controls the assets. The asset owner may or may not be a part of the organization that operates and maintains the assets. Owners directly interact with key stakeholders (e.g., customers, shareholders, regulators, employees and financial agencies), as well as asset managers. The asset owner sets the Goals and Policies (see Figure 2-3) and the Budget for the organization.
2-11
The asset owner should develop and clearly communicate well-defined high-level Goals and Policies and a strategic framework for operating the organization. These goals should be translatable to clear business objectives and measures of performance. Some goals, such as desired internal rates of return, might be very specific and quantifiable. Other goals, such as improving customer perception, may be less specific and incorporated in a long-range plan resulting from a strategic planning process. It is the asset managers responsibility to translate these goals and policies into measurable objectives. Included here may be an outline of organizational roles and responsibilities and business processes that reflect the asset owners policies and philosophies. These goals and policies often start with a mission statement as will be discussed in the next chapter and are used to set performance metrics for subsequent processes. Policy formulation seeks input from various stakeholders, and reflects customer priorities and concerns. Stakeholders can include the asset owners parent company, shareholders, regional operating organizations, state and federal utility, safety and environmental regulators, local governments and the public at large. Each of these groups may have different goals and metrics. Some may be very specific and short term, such as specific earnings per share target or an availability factor. Other goals and metrics may be less well-defined or longer term. These stakeholders (particularly regulators) also may have different constraints on actions that can be taken. These are must do or must not do types of inputs. The asset owner relies on the leadership, vision, values, business objectives, and judgment of the organization and its senior management to establish how to weigh tradeoffs between competing goals and produce a consolidated set of goals and operating policies. Performance Assessment Assessment of asset and system condition and performance provides factual and quantitative information on the performance of the Asset Inventory in meeting established Performance Criteria and information that can be used in subsequent analyses to predict their ability to meet these requirements in the future (an important input for maintenance process planning). Condition assessment and performance monitoring form the basis for management of an asset throughout its life. They provide a basis for adjustments to the various outputs of subsequent processes, ensuring that expected performance goals are met and providing indications where changes are required. Effective asset condition assessment in relation to its service performance is needed to understand and predict the deterioration that leads to reduced asset performance in the future. Assessments of system performance and implementation also may be conducted by external stakeholders (e.g., customer perceptions of the quality of infrastructure condition, or regulator assessment of the provision of services), and these are valid inputs for analysis as well. These evaluations are a key process of PDAM. Understanding the current condition of an asset or the current level of performance provided by a system provides vital information for a series of asset management maintenance decisions and a starting point for predicting future performance. Determining the current level of system performance usually entails straightforward calculations, such as summing the number of customer interruptions. Tracking equipment condition parameters has a number of uses. The most obvious use is for deciding whether some immediate corrective action is indicated. For equipment, many of these evaluations are part of normal maintenance activities. However, it is important that the history of these activities not be
2-12
islanded in the maintenance system. Rather, this information should be available to the larger PDAM process, of which maintenance is just a part. In addition to triggering maintenance, such information should be used to determine how well past asset management decisions have been implemented and whether the expected improvements have resulted. This assessment information also can provide a starting point for projecting future asset or system condition through the use of deterioration models and also to refine existing models. This linkage, and others, will be further explored in the development of the PFM process.
Chapter References
1. Guidelines for Power Delivery Asset Management. EPRI, Palo Alto, CA. 2005. 1010728 2. Publicly Available Specification (PAS) 55: Asset Management Standard, The Institute of Asset Management, London. May 2004 3. Sustaining Local Assets: Government of Victoria December 2003 4. Strategy for Improving Asset Management Practice, AUSTROADS 2002 Sydney
2-13
3
PERFORMANCE FOCUSED MAINTENANCE AND POWER DELIVERY ASSET MANAGEMENT
The previous chapter introduced the fundamental concepts of PDAM. This chapter will expand the explanation of maintenance within PDAM and introduce the concepts of PFM. The Core Maintenance Concept Based on the principles of PDAM, the overriding purpose of a maintenance process is to support the goals and objectives of the asset owner. Many maintenance programs are purely equipment focused. That is, they are designed only around the requirements of the equipment without reference to the larger perspective of how the equipment performance can best support all of the organizations objectives. This narrow focus often puts maintenance programs at a disadvantage when competing for limited resources. The most effective maintenance process is based upon a core mission statement that can be linked to higher-level missions, for example: The purpose of the maintenance process is to achieve the specified levels of asset performance at acceptable costs and in compliance with all safety, health and environmental standards. As required by PDAM, performance targets in terms of asset reliability, availability and service life should be set by direct reference to the business goals. Costs and compliance levels are established similarly. Within this framework, a performance focused maintenance process would have: Work requirements determined on the basis of pre-established standards and criteria for equipment condition or performance. An effective work screening process to ensure all executed work is justified and necessary to support performance goals. A systematic decision process to evaluate and manage risks when selecting maintenance tasks. Execution by the most cost-effective method. Measurement of benefit achieved Performance parameters established, monitored and used as a basis for continuous maintenance and asset performance review and improvement.
The Maintenance Process The only justification for an asset owner to purchase, install and maintain an asset is to receive benefits from the services the asset provides. This is as true for a circuit breaker as it is for a bucket truck. Of course the services may differ, e.g. contribute to safe and effective operation of
3-1
the power system for the circuit breaker, but all service levels and their contribution to the organizations goals should be clearly specified and monitored. The only justification for maintenance is to preserve an assets service level. Therefore, there must be a direct link between the desired performance level and the maintenance objectives. Before exploring the maintenance process, the PDAM process needs to be examined in more detail. As shown in Figure 2-3, assets inventory and the services and performance they provide represent the final output of all the processes of the power delivery organization. These terms are used in the broadest sense, and there are two distinct but related aspects to asset/service performance. The organizations assets are used to provide services to customers end user service levels, often referred to as system performance in power delivery and these can be measured by various metrics. Examples for power delivery include SAIFI, SAIDI, and energynot-delivered and these are developed in the Performance Criteria process. In addition, individual assets or groups of assets are expected to perform at certain levels, which may or may not affect end user service levels. Examples here include equipment failure rates, return on investment, and maintenance costs. The service levels provided with the assets directly impact customers and are directly influenced by Asset condition and operating procedures. The Goals and Policies of the Stakeholders are used to develop criteria for decision-making and measurements of asset and system performance towards achieving those goals. To accomplish this, performance indicators, specific qualitative or quantitative measures that allow performance against a benchmark to be assessed, are required. This step involves transforming the high level strategy into a set of decision criteria for evaluating actions at the lower levels. These criteria could be numerical, such as average restoration time requirements, or they could be qualitative, such as a decision that a safety related task has the highest priority. They range from internal business and engineering indicators to customer and financial perspective indicators. Whatever form they take, the criteria and performance requirements should be constructed to allow the asset and maintenance managers to use a mechanism for prioritizing activities that is consistent with the high-level strategy, goals, and policies of the asset owner. An associated step (not shown), after establishing performance criteria in this process, is risk identification and an assessment of constraints, which define success criteria and unacceptable risk. The impacts from requirements for regulatory compliance are also included here. The Evaluate (Asset Condition and System Performance) process takes inputs from the Identify Gaps process and either directly or with added calculations compares them with desired levels from Performance Criteria. Gaps or anomalies in performance levels may be identified here. The major output is a series of values that reflect the condition assessment of the system, groups of assets, individual assets, or asset components. Examples are outage numbers and durations, number of equipment failures, and cost totals for a specific number and kind of task. This process addresses directly measured performance metrics and values with the objective of determining current asset condition or performance level. Evaluations of asset and system performance provide factual and quantitative information on the performance of the Assets in meeting established Performance Requirements and information to predict their ability to meet these requirements in the future. Performance monitoring, either directly or indirectly, forms the basis for management of an asset throughout its life. It facilitates adjustments to the various outputs of subsequent processes, ensuring that program performance goals are met and providing indications where changes are
3-2
required. Effective asset condition assessment in relation to its service performance is needed to understand the deterioration that leads to reduced asset performance in the future. Evaluations of system performance also may be conducted externally (e.g., customer perceptions of the quality of infrastructure condition, or regulator assessment of the provision of services), and these are valid inputs for analysis as well. The maintenance process is included within the Implement process in Figure 2-3 along with all of the other tactical operations of PDAM. The initial step in developing a maintenance program is to develop maintenance strategies. This process develops maintenance strategies and practices that are to be generally applied in the organization. Based on review and analysis of risk definitions performance criteria and risk assessments, industry practices, manufacturer recommendations and historical data, this processs outputs include maintenance procedures, value and timing of preventive maintenance tasks, condition-based maintenance triggers, maximum operating levels and other similar standards. This process includes reliability-centered maintenance analyses and updating. Of course, maintenance currently is a well-established process in all power delivery organizations. In fact, some confuse good maintenance practices with asset management. Maintenance is an integral part of asset management, but only a part. The objective of an integrated asset management maintenance program is not to simply maintain equipment to some standard but to make certain that maintenance practices are developed in alignment with the organizations goals and objectives. Furthermore, assets are maintained to provide performance levels determined by an evaluation of the required benefits and risk tolerances and an evaluation of alternative actions. The purpose here is to describe how an existing maintenance organization can integrate into the asset management approach and form the basis for PFM. An analysis of these maintenance processes will show that they are simply a special case of the larger PDAM model presented in the preceding chapter. Maintenance Process Diagram A more detailed view of the maintenance process is shown in Figure 3-1.
3-3
Figure 3-1 A Performance Focused Maintenance Process (Source: Guidelines for Power Delivery Asset Management)
The first step is to Identify Important Functions of the power delivery assets that are to be maintained. In addition to the expected equipment design specifications, this process should also consider the risk definitions and performance criteria developed from the asset owners goals. This process aligns the assets performance with the performance level required for the system of which it is a part. This means that identical pieces of equipment could have different performance requirements (e.g. reliability, availability) depending on where they are located in the power delivery system. These different performance requirements should translate into different maintenance requirements.
3-4
Once there is a quantification of the performance requirements for an asset, maintenance tasks and their frequency can be specified to achieve the desired results in the Develop Maintenance Test and Task and Frequency process. Note that these linked processes broaden the maintenance task selection from an equipment decision alone to one integrated into the asset management framework. The next step is Develop Maintenance Program. Since maintenance budgets are limited to some level, there may not be funds for all identified tasks and tests. A program optimization process, similar to that described in Figure 2-3 for the overall PDAM implementation, should be used to insure that maximum benefits are being received from the maintenance program. Perform Maintenance/Test is the next step and it is important that the outputs of this process are available to the overall asset management process. These outputs include the relevant data from equipment history for completed maintenance activities, completed post maintenance test activities, predictive maintenance results, equipment condition data from completed PM, operator rounds, and any other sources of performance data. These data are a part of the maintenance process but they also have uses in developing performance and condition trends, developing aging models for predicting future performance levels, assessing whether past investments have had the desired results and a number of other key PDAM activities. Therefore, there should be a data link to make certain that information that comes from maintenance is available for other uses. One means to facilitate this is to have maintenance and test data reside in the asset information repository, rather than some separate maintenance database. It is also probable that information from non-maintenance activities can be of interest in the maintenance process and there should be a means to transmit this information. Data from both sources should be evaluated to determine whether there is asset degradation. This determination is essentially a subset of the application of the previously described assessment algorithms for maintenance and test data. Here the process is usually much simplified and is often just a comparison of measurement results against asset specifications. An investigation of the asset deterioration may show that a change of maintenance tasks or frequencies would be warranted. If not, then another option may need to be developed (e.g. change in supplier or operating practices). In any case, corrective maintenance will be performed. Equipment Failure is included here as a separate case. Philosophically, a failure could be considered as an extreme case of underperformance and with such a definition the general PDAM model would accommodate failures. In reality, most power delivery organizations treat failures as special cases. The first determination to make is whether or not the failure was preventable. For all cases except run-to-failure equipment, some effort should be made to understand the failure cause. It is recognized, however, that not all equipment failures justify a complete root cause investigation. If no cause can be determined, then the only course is to perform corrective maintenance and continue. Even in these cases, good failure history data should be recorded for possible future analyses, such as developing equipment hazard rate curves. Good failure records will also help to uncover generic problems with specific designs or applications. If the failure was found to be preventable, then the next series of questions are designed to determine whether the failure was maintenance or operations related. If indicated, changes to the procedures or maintenance tasks should be investigated. If no other cause has been identified
3-5
and the risk of another failure exceeds an acceptable level, then a design change should be investigated. PFM Characteristics Outlining the development of a complete maintenance process for a power delivery organization is well beyond the scope of this document and, in fact unnecessary. Maintenance is a wellestablished process in power delivery organizations and significant expertise has been acquired for keeping power delivery equipment in operating condition. On the other hand, maintenance departments are less adept at explaining how to value their contributions to the organizations performance. It is envisioned that PFM will provide both a basis for demonstrating the relative importance of maintenance and a means to improve the effectiveness of existing maintenance programs. Given the opportunity to build a maintenance program from the ground up one would begin by answering a series of questions about each asset starting with: What are the required performance levels for this asset? What is the required reliability? What is the required availability? What is the acceptable service life? What is the acceptable average annual maintenance cost? Furthermore, how does each particular maintenance task affect the assets performance and how is the effect measured? Unfortunately, generally there are no analytical answers to these questions available for power delivery equipment. Rather, the response a maintenance manager would be most likely to receive is something along the lines of: as high as possible, for as long as possible, for as little cost as possible. The direction would be clear but it is difficult to make quantitative decisions without quantitative data. Furthermore, it is difficult to justify maintenance budgets or defend against cuts without an analytical basis. PFM can provide a framework to address the lack of data and begin to provide for quantitative maintenance decisions. In the initial phase, PFM should establish meaningful performance measures that can guide maintenance decisions and track maintenance effectiveness. Moving towards complete PFM requires some steps not always taken in conventional maintenance processes. Understand the operational policies and the service levels demanded of the asset to fulfill the service and policy requirements. Power delivery equipment is not specified in terms of operational reliability. Industry standards may specify some design requirements, e.g. number of operations for a circuit breaker, but there generally is no quantification of equipment reliability or availability. Similarly system and substation designs may be chosen based on computer simulations that calculate an expected service level using nominal component failure rates but there is no link between the design values assumed and the components eventually installed. For older equipment, there is usually some degree of condition degradation but any effect on equipment performance is difficult to quantify. Perform a criticality assessment of each asset. Maintenance resources are limited and their use should be directed to the assets that most contribute to performance goals. Another way
3-6
of stating this is that maintenance should be prioritized to address those assets whose lack of performance has the greatest impact on system performance goals. What may occur as a result of a failure is directly related to where and how the equipment is used in the power system. A measure of the equipments application failure consequences can also be referred to as the equipments criticality. Formally risk is defined as the product of the probability of a hazard causing loss occurring and the consequences of that loss. In the context of this study, the hazard is the failure of the equipment and, as discussed above, ascertaining the probability of equipment failure is difficult. Obtaining some measure of criticality is easier. It is possible to determine the potential for significant failure and consequences - financial, environmental, safety and other adverse impacts on performance goals and rank asset systems in order of their criticality to the organization. Define maintenance requirements based on the equipments risk/criticality analysis. This process will set the most cost-effective maintenance approach for each asset based on the assets contribution to system performance. Select the best failure-countering maintenance strategy for each asset. Use the principles of RCM to choose condition-based tasks, time based tasks or corrective tasks. For defining maintenance requirements and selecting maintenance tasks, consideration should be given to the characteristics of the equipment but also to the specifics of each application and the corresponding operational and environmental stresses. Identical items of equipment may have different maintenance needs based on their different applications. Performance Measurement The recurring theme throughout all of the above discussion is performance measurement. Good performance metrics are important for PDAM and effective maintenance and are absolutely vital for PFM. Properly established maintenance performance metrics provide: A basis for establishing maintenance objectives. A scale to measure maintenance effectiveness Insight into asset performance issues The principles of asset management require that all are linked to the organizations operating goals. The next chapter will discuss performance measures in detail.
3-7
4
PERFORMANCE MEASURES
In a PDAM approach, resource allocation decisions are tightly linked to policy objectives, and based on a quantitative understanding of the results obtained for every dollar invested. This includes maintenance decisions. PFM is a technically sound and justifiable maintenance resource allocation process that depends on the effective use of performance measures. Performance measures allow organizations to translate high-level policy objectives such as service reliability into quantifiable expressions of results to be achieved. They provide information by which managers can make tradeoffs across competing needs and they provide measures of effectiveness in delivery of results. There are three broad classifications for PDAM and PFM metrics: Measures Performance Measures Key Performance Indicators
Background Definitions Before proceeding, it will be helpful to establish some definitions1: Goal - a broad statement of the long-term results needed to accomplish the organizations mission and achieve its vision. Strategic goals are broad statements defining changes the organization hopes to achieve during the strategic planning horizon. Goals focus on outcomes or results and are qualitative in nature. Achievement of strategic goals moves an organization closer to realizing its vision. Goals are defined as broad, ideal future conditions, the results the organization wants to accomplish. As an example, one of the possible goals for distribution could be: Reduce the number and duration of customer outages in every operating region. Strategy - A strategy explains the way a goal or group of goals will be achieved. Strategies are statements of major approaches or methods for attaining each goal and resolving specific issues. An example strategy for the above goal is: Improve the reliability of the ten worst performing circuits in each service area. Objective - Strategic objectives flow directly from strategic goals. They translate the quantitative expectations of goals into specific and measurable targets that the organization should meet to realize the goals. Objectives tell what the organization plans to achieve, who will do it, how to know when it is achieved, and what are the key performance indicators for tracking progress. Unlike goals, all objectives are expected to be met within the planning horizon. As an example, an objective to the above-stated goal to provide the best distribution reliability service in the state could be: Have no feeder with a SAIFI greater than x by y year.
4-1
Some additional definitions that will be useful in the following discussion: Input the amount of resources used to conduct an activity Output product or service; work completed Outcome the results achieved by the output (e.g. customer satisfaction) Target - the specific level of performance the organization is striving to achieve Measure - any meaningful indicator used to determine the magnitude or degree of something. It is important to keep in mind that not every measure relates to performance or is suitable to use as a key performance indicator. Performance Metrics Performance measures are observable, quantifiable measures that align with project and process objectives. They provide the means to track progress toward meeting the objectives. Performance targets are specific values of performance measures that provide the level expected to be attained. This target may be set for a specific time period and with the understanding of a particular level of funding. It provides the bar against which actual performance data will be compared. For example SAIFI is not only a natural unit for measuring service value but also a way to measure performance. There are, in fact, two general classifications of performance metrics of interest to PFM. Results metrics measure what has been accomplished. Process metrics measure how the results were achieved. One conceptual approach to developing a list of performance metrics is to use a process model formulation. Performance criteria defined in the development of the organizations goals can be considered as the outputs of various processes and can be measured. The inputs to the process that produce the result are measured by process metrics. Going back to the SAIFI example, the index is the result of a process with many inputs that can be measured such as vegetation management and equipment maintenance, each of which can be considered as a sub-process with its own process metrics (e.g. maintenance backlog). Defining and measuring performance metrics of both classifications is important for good asset management. It is important not only to direct resources properly to maximize value but also to utilize the resources efficiently and effectively in attaining that value. In addition, good asset managers want to respond to deviations in results metrics quickly. To do this, it is important to know what influences the result metric. Process metrics provide this information. Furthermore, most results metrics are lagging indicators. This is to be expected since they are the output of a process. A distribution manager only knows that the SAIFI metric has gone below target or has not been reached after the fact. Process metrics can be leading indicators (e.g. maintenance backlog). Tracking and managing them can improve the performance of the result measure. For power delivery, there are usually many inputs to a process that has a single result and one should be cautious about developing a performance measure for each. The potential benefit should out-weigh the cost of data collection and storage and it makes no sense to gather information that will not be acted upon. Simply stated, a performance measure is suitable for asset management if it helps the organization make better decisions about where to invest its resources, and if actions taken by the
4-2
organization can influence changes in the value of the measure. In addition to these essential characteristics, the following additional issues should be considered: Does the measure reflect stated policy objectives? Is the measure meaningful to affected stakeholders decision-makers, implementers and customers? Can the measure be monitored with sufficient accuracy and reliability given available resources? Is it feasible to predict the value of the measure under different scenarios? Are output or program delivery measures used to provide early indications of progress towards desired outcomes, and consistency between budget allocations and stated system performance targets? Are procedures in place to ensure that performance information is collected efficiently and accurately, and that results are widely communicated to their intended audiences in a useful and usable manner?
Guidelines for Developing a PFM Measure The selection of a measure depends on the characteristics of the related goal or objective. To determine what measure(s) should be used to monitor progress in achieving a goal or objective, consider the following: Review the goal or objective. What is the desired result conveyed in the goal/objective? Is a particular measure required/mandated? If so, is it adequate or should it be supplemented with additional measures? Review the types of measures: input, output, or outcome. Is the goal or objective about producing a certain quantity of something (i.e. output--how much work was done), is it about operating efficiently (i.e. output/input), is it about how well something was done (i.e., outcome--quality, timeliness, etc.), is it about providing a service to a certain percentage of a given population (i.e. outcome--penetration), or is it about changing behavior (i.e. outcome)? Discuss how will the measure be used in addition to monitoring progress, e.g., to compare to organizations or a national standard, to compare regions within the organization, to demonstrate compliance with a mandate? Who will use the measure? What information is needed to know whether we are making progress toward or have achieved the goal or objective?
Key Performance Indicators Key performance indicators (KPIs) are, in a general sense, the same as the performance metrics discussed above. However, the term has come to be associated with metrics that are tracked and reported to a higher level of authority. For this reason, KPIs usually focus on measuring accomplishments or results. The commonly used reliability indices are an example. They directly measure a dimension of customer service level that is the result of many factors design, maintenance, capital investments, etc. Trigger parameters in performance-based rates can be considered as KPIs and they too are the result of many separate activities.
4-3
Key Performance Indicators are quantifiable measurements, agreed to beforehand, that reflect the critical success factors of an organization. They will differ depending on the organization. Whatever KPIs are selected, they must reflect the organization's goals, they must be key to its success, and they must be quantifiable. Key Performance Indicators usually are long-term considerations. The definition of what they are and how they are measured do not change often. The goals for a particular Key Performance Indicator may change as the organizations goals change, or as it get closer to achieving a goal. Key Performance Indicators help an organization define and measure progress toward organizational goals. Because so many factors can influence a high level KPI, it can be difficult to gain insight when some corrective action is required. For this reason, many organizations develop lower level KPI, for example: maintenance metrics such as number of backlog maintenance orders. In fact, KPIs can be useful at any level of the organization if they are properly chosen. Good KPIs should: Focus on accomplishments rather than activities Utilize readily available metrics Provide meaningful indication of performance to all levels of the organization Promote improvement Allow external comparison (benchmarking) Communicate progress
Distribution Substation Performance Measures In the power delivery industry environment of today, customer service has become a critical issue while cost control remains just as important an element. As customers expect better service at lower cost, regulators are becoming more concerned about service reliability. The number of states enforcing distribution reliability standards continues to increase with more rate decisions tied to service reliability. The challenge for utility engineers and managers is deciding how resources should be spent on reliability to provide the customer the service required, at a price that satisfies all stakeholders. Utilities need to establish reliability and power quality as a core business strategy and take a longer-term view of performance improvement. Being a regulated distribution utility unquestionably requires focused attention on reliability and power quality. Service outages have been shown, in the greater part, to be due to distribution events up to 92% 2 of total outages . Once reliability and power quality have been established as a priority business strategy, performance objectives need to be established. Objectives should be based on performance, not activity. For example, the frequency and width of right-of-way clearing is less important than the effectiveness of vegetation management operations as reflected by system reliability.
Customer service is of primary importance to any distribution organization and a number of indices have been established (IEEE 1366-2003) 3. Load point indices measure the expected number of outages and their duration for individual customers. System indices such as System Average Interruption Duration Index (SAIDI) and System Average Interruption Frequency Index
4-4
(SAIFI) measure the overall reliability of the system. These indices can be used to compare the effects of various design and maintenance strategies on system performance. One major issue with reliability metrics is the debate about comparing these indices from one geographic area to another and exactly how the input data should be applied in making the 4 calculations . In addition, there are concerns about how to normalize the indices for adverse weather. Many state public utility commissions require utilities to compute and track certain reliability indices, but comparing them from region to region and utility to utility has been problematic due to differences in how the data is applied, system designs, weather differences, and operating requirements. Because of this, the indices are limited in their usefulness. If the calculation method is kept the same, they are useful within a specific geographic area in evaluating changes in reliability over time and as a measurement of the effectiveness of maintenance practices. Nonetheless, these indices are well established and several are listed below: Sustained Interruption Indices SAIFI System average interruption frequency index SAIDI System average interruption duration index CAIDI Customer average interruption duration index Momentary Interruptions MAIFI momentary average interruption frequency index Load Based Interruption Indices ASIFI Average System Interruption Frequency Index ASIDI Average System Interruption Duration Index
Due to their widespread use, there is no need to discuss these indices further here. Suffice it to say that these could all be considered KPIs. Some example KPIs used by power delivery organizations are listed below. This list is not intended to be exhaustive but rather to show the variety of possible indicators that can be tracked. The KPIs are grouped by reliability, safety, work processes, cost, equipment, and financial. Reliability ASAI CAIFI CAIDI LAIFI MAIFI SAIFI SAIDI
4-5
Number of events without switching remedy Failures without switching remedy / Failures total Number of customers impacted by outage Number of customers impacted by distribution bus outage events Correct performance of protective equipment per event Number of distribution bus outage events Load not served Vegetation caused SAIFI Vegetation interruptions per 100 miles Distribution system vegetation management non-storm SAIFI contribution Safety Number of human errors events Man-hours without injury Lost workday case incident report OHSA reportable incident rate: by department Responsible vehicle accidents: by department Safety required training: by department Work Processes O&M costs per bay by voltage % Critical maintenance complete Restoration time Maintenance backlog PM hours/emergency hours CM work orders/ PM work orders CM resulting from PM inspections Non-weather overtime as a % of regular time: by department Overtime hours by reason Overtime hours by reason as a % Distribution system routine trimming program scheduled vs. completed miles
4-6
Jobs started on time: by department Work completed on time: by department Event investigation performance: by department Mapping & document services: work backlog Planned/completed as scheduled Costs Cost of maintenance By equipment group By equipment model / manufacturer By area / region By maintenance category
PM/CM cost ratio O&M costs / MWh delivered O&M costs / MW installed Call Center Costs average / customer Distribution system trimming program cost per mile O&M cost/customer Capital costs/customer On hand inventory $/system Mw installed On hand inventory $/customer Planned O&M cost/emergency O&M cost O&M cost/ Mw installed O&M cost/ MWh delivered Capital cost/ Mw expanded Capital cost/ customer connected Total budget performance by department O & M budget performance by department Capital budget performance by department Spending on consumables, small tools and material: by department
4-7
Sf6 Gas Use - number of bottles per month Equipment Average equipment age % Tasks requiring custom-built parts Average equipment utilization factor Failure rate by equipment group Mean Time Between Failures - by equipment types, service areas Mean Time To Repair - by equipment types, service areas Substation SAIDI Substation SAIFI Pole replacement miles inspected Financial Capacity investments/Capacity installed [$/MW] Capital investment/Total book value Book value / MW installed Book value / MWh delivered Inventory turnarounds Inventory costs/MW installed Inventory carrying charge
PFM and Performance Measures PFM can help to identify performance measures that can be used to make resource allocation decisions with respect to maintenance and operation, and develop a framework that maintenance managers can use for selecting suitable performance measures and setting performance targets. At this stage of the development of PFM, rather than attempting to recommend a single set of performance measures suitable for any power delivery organization, it is more productive to
4-8
define criteria for what constitutes a useful performance measure within the context of asset management and PFM. While the choice of specific performance measures is important and can influence what gets done, the ways in which performance measures are used and integrated into maintenance processes deserve equal attention. While utilities face similar challenges of how to make the best use of available resources, there are significant variations across companies with respect to service territory, policy objectives, organizational structure, system design, management processes, staff capabilities, data and performance measures already in place. Each of these factors has a bearing on what kinds of performance measurements and approaches will be most successful in a given situation. Without question, the paramount performance measures for distribution systems are the wellestablished reliability indices, SAIFI and SAIDI. Although the details of their calculation may differ, most distribution organizations track performance in these terms. More recently, several states have incorporated reliability indices into their regulatory structure and linked them to rates. Such performance based rate mechanisms make it easier to monetize the value of a reliability index but, as previously discussed, these indices are output measures. They are the result of the combination and interactions among a large number of variables. Asset and maintenance managers only can affect these output performance indicators by changing one or more of the controllable input variables. The problem is that the relationship between changes in the inputs and the outputs are not well quantified. For example, it is clear that customer reliability can be influenced by substation equipment reliability and it is possible to determine how many customers would be impacted by a particular equipment failure by examining system topology and layout. The problem comes when trying to determine how equipment maintenance can improve a customer reliability index. What would be the corresponding improvement in customer reliability for increased equipment maintenance? Without quantifiable relationships, it is difficult to direct resources most effectively. This issue will be further explored in Chapter 5. Secondary KPIs Figure 4 -1 shows how both power quality and the number of interruptions influence a higherlevel goal, reliable customer service. Forced interruptions are tracked by a number of previously reviewed reliability KPIs. Managing these indices is an important responsibility for distribution system operators but, as discussed earlier in this chapter, these are lagging, or output, KPIs. They only indicate a deviation from the desired performance after it has occurred. To better anticipate and control possible shortfalls in performance, one must identify and track the process inputs, the leading indicators. For the case illustrated in the figure for example, some input factors to the maintenance process that could affect outage duration include: Crew productivity The Crew assignment process Crew skills and knowledge Availability of replacement parts
4-9
Of course, there are many other factors and these four are often tracked as internal processes within a maintenance organization. However, their impact on higher-level goals such as customer satisfaction often is not considered or quantified. Without an explicit link, it is more difficult to justify reallocation of resources or manage proactively. PFM starts by examining all the factors that influence a performance goal, then developing a relationship measure and finally setting performance targets. An illustrative but simplified example of the process might start with the question: What determines customer service reliability?
4-10
Focusing on maintenance, what are some secondary maintenance KPIs that can be tied to reliability? A partial list would include such measures as: Preventive maintenance backlog PM/CM cost ratio Vegetation incidents by line PM by equipment Equipment condition index Age Diagnostic tests results Time since last overhaul
The next step in the PFM process would be to develop a set of management metrics to measure performance that would combine a number of secondary KPIs in order to provide a broad based leading indicator. The details of the process would depend on the organizations higher-level performance goals and metrics. Examples include: Corporate objectives or mission statement Balanced score card Incentive programs Performance based rates
Links to the next level of KPIs would be established and then the secondary KPIs would be identified. The types and quality of data would drive this process available. Depending on the particular activity, tertiary KPIs may be established. Once the linkages were determined, influence models would be developed to quantify the relationships among the process inputs and the final output, the higher level goals. The entire process would start with simple models using available information expand with experience. The PFM models would: Show where the organization is now Quantify and track changes in higher level KPIs as changes are implemented in lower level processes Compare areas, stations, equipment types to highlight good and bad actors Provide analytical support for changes in resource allocation Give people meaningful targets to which to manage
4-11
Chapter References 1. Guidelines for Power Delivery Asset Management Long-range and Strategic Planning. EPRI, Palo Alto, CA: 2006. 1012496. 2. Distribution Reliability Indices Tracking within the United States. EPRI, Palo Alto, CA: 2003. 1008459. 3. IEEE 1366-2003: IEEE Guide for Electric Power Distribution Reliability Indices. New York: Institute of Electrical and Electronics Engineers, 2003. 4. ORNL/TM-2004/91 MEASUREMENT PRACTICES FOR RELIABILITYAND POWER QUALITYA TOOLKIT OF RELIABILITYMEASUREMENT PRACTICES. Oak Ridge National Laboratory. June 2004.
4-12
5
PERFORMANCE FOCUSED MAINTENANCE FOR DISTRIBUTION SUBSTATIONS
Maintenance is only one of the activities affecting distribution reliability. Geography, circuit design, protection schemes, restoration logistics, equipment specifications, and load characteristics all play important roles in determining the level of service provided. The IEEE 493-1997 Standard1 provides more reliability considerations for distribution systems. For the most part, distribution equipment maintenance programs have been focused on component reliability without any well-established, quantified link to the requirements for and impact on system performance and customer service. In order to apply PDAM to distribution maintenance, the description of distribution system equipment reliability as a function of component operating age and maintenance history is a fundamental but unsolved requirement. Distribution Substation Maintenance It is recognized that there is no industry standard for the designation Distribution Substation and utilities have different boundaries as to what constitutes their distribution and transmission system. Therefore, the scope of this work will be limited to substations at approximately 50 kV and below. Utility companies have many different definitions as to what constitutes their distribution and transmission systems. The upper boundary varies considerably while the lower boundary is usually the customer meter. Therefore, while some utilities will consider high voltage equipment as part of the distribution system and others will not, every utility has lower voltage equipment in its distribution system. It is this lower voltage substation equipment, generally considered by all to be distribution that is the focus of this project. In most technical discussions, distribution equipment voltage classes are usually considered to be below 50 kV although transformers with greater high side voltages are often included in distribution substation maintenance programs. Maintenance is often organized on three levels: Maintenance plans are directed at specific equipment types, makes and models and are composed of tasks specific to that equipment Maintenance programs combine sets of maintenance plans and are used to organize and manage maintenance resources Maintenance strategies link maintenance programs to the organizations goals and objectives.
Performance measures, as discussed in the preceding chapter, can be applied at all three levels but PFM is primarily focused on the interface between maintenance programs and maintenance strategies. Maintenance programs generally are constrained by resource limitations, primarily economic. On the other hand, there are two main drivers for maintenance:
5-1
Safety. The primary consideration in any maintenance program is the safety of the public and utility personnel. The safety driver is the most difficult to quantify. The two questions that need to be asked are: What is the acceptable risk of a harm-causing component failure and how does one know when the risk has been reduced to acceptable levels? Utility worker safety translates into maintaining circuits and components so that the utility worker can operate them safely in both emergency and non-emergency situations. Service Reliability. Service reliability is reflected in the frequency and duration of customer interruptions. The contribution that maintenance makes to interruptions must be carefully analyzed in concert with reliability targets and other activities that impact service reliability in order to develop the lowest cost solutions.
Ideally, to address the drivers effectively while working within the constraints, a maintenance process, which crosses all three levels, should: Plan equipment maintenance taking into account how maintenance interacts with other activities and events that impact service performance Be based on equipment reliability criteria and target indices for local service areas and circuits that provide a fair balance between reasonably achievable circuit performance, life cycle management and customer expectations. Direct resources towards maintenance work that will have the maximum impact on meeting corporate goals.
In reviewing a maintenance program, these questions present themselves: What is to be accomplished? Short-term goals Long-term goals How is progress measured? Metrics KPIs What is needed? For goals For measurements What is available? Data Analysis tools Decision support tools How to fill the gaps?
When decisions are made about the level of reliability to be achieved in a system, attention must be paid to the economic factors, such as the incremental costs of reliability, the benefits expected
5-2
from a change in reliability, and the allocation of the reliability investment among the system components. PDAM and Distribution Maintenance Based on the principles of Power Delivery Asset Management presented in Chapter 2, an approach to distribution maintenance optimization would: Define management goals and objectives, including risk criteria and risk elements. Translate goals and objectives into performance targets. Identify and analyze internal and external parameters with significant impact on critical equipment service life and failure rates (e.g.. design, age, repair history, loading history) Develop average fault statistics (fault rates, repair times) Assess maintenance and repair costs for critical components for potential failure scenarios. Determine customer outage costs for potential failure scenarios along with any consequential costs such as regulatory penalties. Develop a program based on maximizing cost/benefit and available funding. Implement the least cost plan meeting the required service levels.
From the above, it is clear that equipment asset management must be approached from a system performance point of view, and investment decisions should be coordinated with operation and maintenance decisions. To be most effective, application of asset management techniques should be based on knowledge of the condition of the components of the network. However, due to the high number and relatively low costs of the installed equipment base in distribution systems, individual condition monitoring is not practical. Furthermore condition assessment of the components using diagnostics would be time-consuming and very costly. Therefore, the only practical way to implement an asset management approach to distribution systems is to use statistical methods. Simply stated then, optimal maintenance would provide just the required system performance level within the specified risk and cost boundaries. Since system performance is a function of equipment reliability then it follows that optimum maintenance would be that which provides the required level of equipment reliability. Determining the required level of equipment reliability to satisfy system performance levels is beyond the scope of this work but many techniques exist for designing and calculating expected distribution circuit reliability. The issue for this project is the linkage of maintenance to expected equipment reliability. The challenge then is to develop the relationship of distribution system equipment reliability as a function of the component operating age and maintenance. PFM Fleet Management One of the most pressing needs for new and better tools to support power delivery maintenance, including replacement, is in the area of aging asset management. Even for those utilities not faced with a large aging asset problem, quality long-range planning requires a good understanding of the current and projected future condition of the asset base.
5-3
It is well accepted that the risk of equipment condition deterioration and wear-out failures increase as equipment approaches the end of its useful service life. The rationale and timing of increased maintenance or investment decisions in anticipation of this increased risk have been traditionally left to historic patterns and engineering judgment. This may result in higher costs to the utility and its customers if investments are not optimally timed, i.e. if made too early, they may result in higher carrying costs; if made too late they may result in reduced service and higher failure costs. PFM with an aging asset infrastructure makes it increasingly critical to identify equipment hazard functions of major asset classes and to understand the risks and influence of critical variables on equipment failure rates. With better failure rate information, risk based approach models of the underlying distribution of failures or bath-tub curve of each asset type can be developed to allow optimizing the risk-cost function in the maintenance planning and decision making processes.
Numbers of Units
Ageing
80
70
60 45 30 15 5 1
25 10
Age
Figure 5-1 A Hypothetical Equipment Age Distribution and Failure Rate Curve
Historically, projected equipment failure rates were computed by dividing the number of past failures by the equipment-years considered in the studies. These historical failure rates were assumed to be constant throughout the equipment life and used for replacement and spares planning. The general situation is depicted in Figure 5-1. A histogram showing the number of units of a particular equipment type in several age brackets has been superimposed on a failure rate curve that shows an increasing failure rate with age for that type. The age profile of the units reflects previous investment patterns. As time progresses, an increasingly larger percentage of the equipment population will move into the range of higher failure rates. This type of histogram gives rise to the term asset walls. A significant concentration of a particular asset in a group of adjoining age brackets looks similar to a wall moving forward in time. As illustrated in Figure 5-1, the assumption of constant failure rate only applies if the equipment in question is operating in the flat portion of the hazard rate curve. Furthermore, many factors can influence equipment performance causing variations in equipment failure rates with time and usage. These could include equipment age, loading, manufacturer and maintenance history. A better understanding of how these additional stress factors affect equipment performance is required to correlate the relationship of various parameters with the probability distribution of equipment time to failure and failure rate.
5-4
Approaching asset walls and new operating requirements make projections based on history increasingly risky. Planning maintenance, estimating future capital requirements and making the best effective capital funding decisions in the present requires a better understanding of asset performance projections over the long term. Understanding individual asset performance is important for tactical planning but understanding the collective performance for groups of asset types, often called fleets, is important for PDAM and distribution PFM, where maintenance programs are usually not tailored to individual units. Linking Maintenance to Equipment Reliability Business pressures to contain costs while preserving service standards, including the spreading use of performance-based rates, have highlighted the lack of understanding of a quantifiable relationship between maintenance and equipment reliability. Intuitively, maintenance should improve equipment reliability and consequently system performance but it is difficult to quantify any improvement. However, equipment failure rate models are required for power delivery organizations to plan, operate and maintain their systems at the required levels of reliability for the lowest possible cost. Managers would like to know what the increase in reliability would be for an incremental increase in maintenance funding or conversely, what is the impact of a reduction in maintenance on system performance. This topic has become the focus of a number of research efforts. As part of this project, a literature search of pertinent studies was conducted and three promising approaches were identified. All three are in developmental stages and have been applied only to limited degrees. Nonetheless, all offer a possible approach for establishing a method for modeling the relationship between maintenance efforts and equipment condition and reliability. This linkage is necessary for implementing a complete performance focused maintenance process. The three approaches represent the current state of the art in this area and are briefly discussed below. Risk-based Planner for Asset Management 2 Maintenance decision analysis is a combination of engineering and financial analyses and requires complex mathematical modeling. However for practical application, any software must be user-friendly to make the operation of the model simple. With this in mind, substantial effort has been put into developing a decision-support tool to support Risk-based Asset Management 3 (RiBAM) starting in the late 1990s . Anders et al have developed the asset management planner (AMP) model as a mathematical model and an associated computer program designed to calculate the expected remaining life to failure of equipment. The expected times to failure from various life stages of the equipment are used to construct the life curves required by RiBAM. The basic idea in the AMP model is the probabilistic representation of the deterioration process through discrete stages. There are three major factors that contribute to the aging behavior of equipment: physical characteristics, operating practices, and maintenance policy. The physical characteristics can be summarized as: the type of construction, the power rating, and other equipment physical and design factors. Operating practices can significantly impact the deterioration process. The maintenance policy affects the operating state of the equipment by affecting its condition. The three components to a maintenance policy that the RiBAM approach recognizes are monitoring or inspection (how is the equipment state determined), the decision process (what determines the
5-5
outcome of the decision), and finally, the maintenance options (or possible decision outcomes). Component state identification is accomplished through the use of scheduled inspections. Decision events generally correspond to inspection events, but can be triggered by observations acquired through other sources. In practice, the decision process will be affected by what state the equipment is in, and by external factors such as economics, current load level, forecast load level and other such factors. The AMP is a mathematical model designed for equipment exposed to deterioration but undergoing maintenance at prescribed times. It computes the probabilities, frequencies and mean durations of the states of such equipment. The basic ideas in the AMP model are the probabilistic representation of the deterioration process through discrete stages, and the provision of a link between deterioration and maintenance (Figure 5-2). Clearly, maintenance is expected to slow the rate of deterioration. In most applications, it is sufficient to represent deterioration by three stages, an initial (D1), a minor (D2), and a major (D3) stage. This last is followed, eventually, by equipment failure (F) that requires extensive repair or replacement.
Figure 5-2 Equipment Deterioration Is Modeled in a Markov Process (Source: Risk-based Planner for Asset Management)
In order to slow deterioration and thereby extend equipment lifetime, maintenance will be performed according to some pre-defined policy. In the model of Figure 5-2, regular inspections (I) are performed which result in decisions to continue with minor (M) or major (MM) maintenance or do nothing. The expected result of all maintenance activities is a single-step improvement in the deterioration chain; however, allowances are made for cases where no
5-6
improvement is achieved or even where some damage is done through human error in carrying out the maintenance resulting in the next stage of deterioration. The choice probabilities (at the points of decision making) and the probabilities associated with the various possible outcomes are based on user input and are estimated from historical records. The following is a complete description of the notation used in Figure 5-2. D1, D2, D3 denote deterioration states with D1 denoting initial, D2 minor and D3 major deterioration conditions. F denotes equipment failure to perform its function due to deterioration M21, M22 and M23 denote minor, medium and major maintenance activities associated with the deterioration state D2 and M31, M32 and M33 are corresponding maintenance activities associated with the deterioration state D3. I1, I2 and I3 represent the inspection activities associated with the deterioration states D1, D2 and D3, respectively. The decision whether and what type of maintenance to perform is made following an inspection.
All of the above assumptions are incorporated in the program in an appropriate state-space Markov model. It consists of the states the equipment can assume in the process, and the possible transitions between them. In a Markov process the rates associated with the transitions are assumed to be constant in time. Mathematically, this model can be solved by well-known procedures. The solution yields all the state probabilities, frequencies and mean durations. Another solution technique, employed for computing the so-called first passage times (FPT) between states, will provide the average times for first reaching any state from any other state. If the end-state is failure, F, the FPTs are the mean remaining lifetimes from any of the initiating states. The first passage times are used to generate the life curve of the equipment. The life curve is constructed by associating the remaining times before failure computed by the program with the description of the deterioration states provided by the user. Also, using Monte Carlo simulation, the probabilities of failure in a specified time horizon can be computed. The present value approach is employed in this model to formulate maintenance decisions on aging equipment include timing, and the time value of money. In selecting the best course of action, the proposed alternatives are compared with some reference action. The corresponding cost difference is referred to as the net present value (NPV). In the case of maintenance, the NPV can be obtained for several re-investment options and are compared with the present maintenance policy. Cost computations involve calculation of the following cost components: Cost of maintenance activities, Cost of the action selected (overhaul or replacement), Cost associated with failures (cost of repairs, system cost, penalties).
To compute the PV, inflation and discount rates are required for a specified time horizon. The time horizon is a period of time, starting at the present and ending after a chosen number of
5-7
years, for which the costs of the various operating and maintenance options are calculated and compared. The cost of maintenance over the time horizon is the sum of the maintenance costs incurred by the original maintenance policy for the duration of the delay period, and the costs incurred by the new policy for the remainder of the time horizon. The costs associated with equipment failure over the time horizon can be computed similarly except that the failure costs before and after the action must be multiplied by the respective probabilities of failures, and the two products added. The cost of maintenance and failure with the present maintenance policy is the potential benefit of the decision, that is, the consequences avoided by taking new action. Since there are many uncertainties related to the prediction of equipment life, probabilistic analysis is the best tool that can be applied. The objective in optimizing the model is to determine the type of maintenance action and the year this action is taken so that NPV is maximized while not exceeding constraints such as budget limits or forced outage rate limits. In other words, the goal is to optimize NPV while satisfying financial and reliability constraints. The program uses the concept of a life curve and discounted costs to show the effect of equipment aging with time. Usually, the decision-maker has one of the following options: Continue as before Do nothing Refurbish Replace
Each of these actions will have a different effect on the shape of the life curve during the period following the action. The curves generated by the program are based on complex Markov models. To generate life curves automatically, a default Markov model for the equipment has to be built and stored in the database. The RiBAM model has the advantage of being incorporated into a user-friendly, beta grade program. It provides for a complete analysis inasmuch as the program considers equipment condition, maintenance actions and costs. On the other hand, to utilize the full benefits of the models capabilities, a fair amount of input data is required. It is not clear how many different equipment models would be required to represent a typical distribution substation or whether sufficient historical data would be available. Empirical Life Models
4
The work referred to here is a new method that is predictive and uses a derived life model based on material degradation, a probabilistic failure model and the enlargement law to predict the probability density of failure and the failure rate of the electrical components of a distribution system. Typical aging processes of electrical components are considered to be due to electrical, thermal and mechanical processes. Therefore the aging processes are factored into an equipment life model. The method has shown good results with some preliminary application to German networks. The fundamental principle behind the modeling is that equipment failure is strongly correlated to the aging of the electrical components, where failure rates are not constant for the whole lifetime.
5-8
Therefore modeling the aging phenomena of the materials of the electrical components is integral to the development of the applied models for the assessment of the condition of the electrical equipment. The rationale is that due to the presence of degradation stresses, such as electrical, thermal, mechanical and ambient temperatures, the aging of the materials in the equipment is the dominant contributor to equipment failures. Therefore it is necessary to consider the material characteristics in order to assess the service life of the electrical components. The determination of the electrical, thermal and mechanical degradation of electrical components in this study is limited to the establishment of an empirical correlation. It is postulated that this, with a minimum of adjustable parameters, can successfully predict the degree of degradation of electrical equipment under the influences of electrical, thermal and mechanical stresses. If a generic combination of electrical and thermal stresses is applied to an electrical component, a suitable life function t of the electrical component can be established according to the following correlation:
t = t0 ( E/E 0 )
where:
-BT
, T = 1/0 - 1/
n is the voltage endurance coefficient, E 0 the lower limit of the electrical stress (below which the electrical aging can be neglected), B proportional to the activation energy of the main thermal degradation reaction, T the conventional thermal stress, the absolute temperature, 0 a reference temperature, and t0 the corresponding lifetime at the electrical stress E0 and the temperature 0. The proposed electro-thermo life model comes from a combination of two single-stress models, i.e. an electrical life model and a thermal life model, which are represented by the InversePower-Model and the Arrhenius-Model respectively. This is done by assuming that the aging rate under two combined stresses is the product of the aging rates under each single stress. If a mechanical stress is considered, the mechanical life model is equivalent to the electrical one, which can be expressed by the Inverse-Power-Model. As both electrical and thermal stresses influence each other, the combined electro-thermal life model will result in an overestimation of the synergism between stresses, which would lead to an underestimation of life especially at high stresses. Therefore a suitable corrective function is introduced. It is assumed that an accepted statistical model for determining the likelihood of failure at a given stresses would be consistent with the shape of an exponential function and that such an
5-9
empirical function therefore could be described well by the Weibull-function. The twoparameter Weibull-function used can be written as:
p ( E) = 1 exp [- ( E/E63%) ]
where: p is the function of failure probability of an electrical component, is the shape parameter, which can be obtained by breakdown tests, and E63% is the electrical withstand strength of the Weibull-function, whose corresponding failure probability is 63%. As a multi-stress model the Weibull-function provides a probabilistic failure model giving the failure percentiles for each pair of stresses. Finally, assuming the non-dependence of the failures, which take place in parallel with respect to space (volume-effect) and time (time-effect), the enlargement law is applied to scale up the model. This model is simple in its construction and directly relates to actual physical degradation. It produced good results for overhead lines, cables, circuit breakers and transformers. However, it does require some material parameters that may not be readily available and it appears that German utilities have more detailed and complete histories of distribution equipment failures than can be assumed for the typical United States utility. Finally, this approach does not directly include maintenance actions in the model. The effects of maintenance would have to be accounted for by changes in one or more of the model material parameters. Nonetheless, it may be possible to use this approach to develop empirical relationships if sufficient data were available. Failure Rate Modeling Using Equipment Inspection Data 5 This paper presents a method to overcome the inherent limitations of assuming average component failure rates by proposing a method to estimate component specific failure rates. It uses equipment inspection data to assign relative equipment condition rankings and these rankings are then mapped to a failure rate function based on worst-case units, average units, and best-case units. The paper goes on to recommend failure rate models for several equipment types and presents a method of calibration based on historical customer interruptions. The proposed methodology starts by assuming that there are available condition index values for the components of interest. The condition score is the result of some combination of inspection results where the best inspection outcome is given a value of 0, an average inspection outcome is given a value of , and the worst inspection outcome is valued at 1. Each inspection item result is assigned a weight based on its relative importance to overall equipment condition. The final condition of a component is then calculated by taking the weighted average of all inspection item results. Then, by definition, a weighted average of 0 corresponds to the best possible condition score, a weighted average of corresponds to average condition score, and a weighted average of 1 corresponds to the worst possible condition score.
5-10
After exploring a variety of mapping functions, the authors have empirically found that an exponential model best describes the relationship between the normalized equipment condition and equipment failure rates. The specific formula chosen is:
f (x) = A e + C
where: f = failure rate x = condition score There are three unknowns, the parameters A, B, and C, and therefore three data pairs are required to solve for them. The previous section has developed a condition ranking methodology that, by definition, results in best, average, and worst condition scores of 0, 1/2, and 1, respectively. Therefore, these three condition score should correspond to f (0), f (1/2) and f (1) respectively. Since a condition score of is by definition an average condition, f (1/2) can be approximated by taking the average failure rate across many components or by using average failure rates documented in relevant literature. The other two failure rates are more difficult to determine but the authors suggest that they can be derived through benchmarking, statistical analysis, or heuristics. The parameters A, B, and C can be solved for once these three failure rates are set. The paper provides a number of failure rate model parameters and failure rates for various equipment types based on past studies found in the literature. The authors caution, however, that these values only should be considered as starting points. The next step in the modeling process is to calibrate the model to the system under study. The focus of the paper is to use the equipment models to model the level of service provided by a specific distribution system. Therefore, after creating a system reliability model, it is proposed that the component reliability data be adjusted to minimize the differences between the predicted system reliability and the historical system reliability by calibrating the condition-mapping parameters so that a distribution-based error function is minimized. Three possible error functions with three levels of granularity are suggested: Individual customer reliability, Histograms of customer reliability, Statistical measures of customer reliability. The paper concludes by presenting an example application of the modeling process. Although the paper presents the equipment failure rate model primarily as a contributor to system modeling, it has potential value for modeling the effects of maintenance on equipment reliability. Since the model uses equipment condition as an input, maintenance impacts on equipment condition would be translated into changes in equipment reliability. It seems that it should be possible to use equipment condition indices developed through means other than onsite inspections in this methodology as well as the condition score presented. Combined with other EPRI work on maintenance triggers6 the modeling approach presented in this paper may provide a way to better characterize and track maintenance effectiveness. Still, the question of the availability of suitable and sufficient historical data remains.
Bx
5-11
Practical and reliable distribution equipment failure rate models are still in the early stages of development. The need for such models will only grow as power delivery organizations struggle with demands for quality service at low prices and the uncertainties posed by an aging asset base. Fleet Management One of the most pressing needs for new and better tools to support power delivery maintenance, including replacement, is in the area of equipment asset management. Even for those utilities not faced with a large aging asset problem, quality long-range planning requires a good understanding of the current and projected future condition of the asset base. It is well accepted that the risk of equipment condition deterioration and wear out failures increase as equipment approaches the end of its useful service life. The rationale and timing of increased maintenance or investment decisions in anticipation of this increased risk have been traditionally left to historic patterns and engineering judgment. This may result in higher costs to the utility and its customers if investments are not optimally timed, i.e. if made too early, they may result in higher carrying costs; if made too late they may result in reduced service and higher failure costs. PFM with an aging asset infrastructure makes it increasingly critical to identify equipment hazard functions of major asset classes and to understand the risks and influence of critical variables on equipment failure rates. With better failure rate information, risk based approach models of the underlying distribution of failures or bath-tub curve of each asset type can be developed to allow optimizing the risk-cost function in the maintenance planning and decision making processes. Historically, projected equipment failure rates were computed by dividing the number of past failures by the equipment-years considered in the studies. These historical failure rates were assumed to be constant throughout the equipment life and used for replacement and spares planning. The general situation is depicted in Figure 5-1. A histogram showing the number of units of a particular equipment type in several age brackets has been superimposed on a failure rate curve that shows an increasing failure rate with age for that type. The age profile of the units reflects previous investment patterns. As time progresses, an increasingly larger percentage of the equipment population will move into the range of higher failure rates. This type of histogram gives rise to the term asset walls. A significant concentration of a particular asset in a group of adjoining age brackets looks similar to a wall moving forward in time. As illustrated in Figure 5 -1, the assumption of a constant failure rate only applies if the equipment in question is operating in the flat portion of the hazard rate curve. Furthermore, many factors can influence equipment performance causing variations in equipment failure rates with time and usage. These could include equipment age, loading, manufacturer and maintenance history. A better understanding of how these additional stress factors affect equipment performance is required to correlate the relationship of various parameters with the probability distribution of equipment time to failure and failure rate. Approaching asset walls and new operating requirements make projections based on past history increasingly risky. Planning maintenance, estimating future capital requirements and making the best effective capital funding decisions in the present requires a better understanding of asset performance projections over the long term. Understanding individual asset performance is important for tactical planning but understanding the collective performance for groups of asset
5-12
types, often called fleets, is important for PDAM and distribution PFM, where maintenance programs are usually not tailored to individual units. Power Transformer Fleet Management One of the approaching asset walls of most concern is that of power transformer. These are obviously critical operating components in the power system, are often the most expensive, and require long and disruptive repair and replacement procedures. Running all power transformers to failure may entail unacceptable financial and operating risks. Because of the skewed demographic distributions common in many utilities and the fact that significant numbers of transformers may be at the back end of the failure rate bathtub curve, existing methods need improvement to provide informed long-range planning for effective management of this aging transformer generation. In response to this need, EPRI launched a project to help asset and maintenance managers deal with the problem of managing populations of aged transformers by formulating innovative methodologies to analyze transformer investment strategies. The results illustrate a successful application of the integration of equipment asset management with corporate, financial asset management. This work is reported in Innovative Methods for Managing Aged Transformer Fleets7 and follow-on efforts are summarized below. The latter work was directed at practical applications of the methodology for a specific set of circumstances and included: Analysis of several possible power transformer maintenance and replacement strategies Projections of the numbers of failures going forward for a twenty year planning horizon, Net present value comparisons of these potential strategies, Budgetary requirements for each of these scenarios for the planning period.
The results discussed below draw from that work and demonstrate the types of methodologies and data that could be extended to distribution equipment and would be supportive of a PFM process for transformers and other equipment types. In the referenced project several innovative methodologies were analyzed and formulated for the types of scenarios and business case studies that typical utility asset and maintenance managers need to consider. Integral to the development of such business case analyses is the ability to project the rate of failure of the transformer population at risk. In the basic case, this is calculated by convolving the hazard rate function with demographic data as previously illustrated in Figure 5-1. The convolution is the sum of the products of the number of transformer units in each age bin times the value of the hazard rate function for that specific age bin. The hazard rate function is fixed for a given population; however for each year or interval into the future, the demographic distribution moves to the right, causing more overlap and higher numbers of projected failures. The key to successful application of this approach is to establish the proper hazard rate function. As previously discussed, it is not a straightforward task to generate an accurate curve for the assets of interest. Quantitative data on the hazard rate function for transformers are sparse. The reliability performance of a transformer is first determined by the physical characteristic of the transformer as set by its design, manufacturer, and materials. Hazard rate is further modified by the stresses experienced by the transformer. Loading, switching transients and environmental factors all stress the unit. Finally, maintenance practices may influence the condition of the
5-13
transformer and hence its ability to withstand stresses. The appropriate hazard rate function can be derived from actual failure data from the utility (if there is enough available) or, for the normal aging portion at least, from a model for aging based on standardized methods used in ANSI C57.91 or IEC standards. Such well-established models relate hot spot temperature to life expectancy as a function of oil condition. Based on such a model, the hazard rate function can be calculated for specific loading conditions and insulation system condition. Note though that this is only one of the contributing factors to the overall hazard rate. Described below in a later section is additional EPRI work to address the scarcity of data for constructing representative equipment hazard rate curves. For the focus of this transformer study, sufficient utility failure data was available to construct a hazard curve for the transformer type of interest. A detailed analysis, graphical analysis of multiply censored data, which takes into account all of the data from a population, namely the failure data and data associated with units in the population which have survived to date, was used and several conventional statistical distribution functions tested. The Normal distribution provided the best fit to the available data. The transformer owner used an economic value added (EVA) model as a tool for evaluating company performance and investments. On an overall company assessment basis, EVA is a value-based financial measure, which reflects the amount of shareholder value created or lost on an annual basis. It has become a tool of choice for controlling operations in some companies and for evaluating alternative financial investments. This tool provides decision-makers a logical and convenient method to evaluate the benefits of deferring investments, as offset by the influence of inflation, and to provide management business case analyses which include projected budget variances, earnings measures such as ROI and EPS, and discounted cash flow measures such as projected NPV of the cash flow for selected discount rates. Other tools and methods are available to accomplish the same function. The financial fundamentals of EVA were adopted for the projections of the economic consequences of candidate fleet strategies in the study. To consider business risk, probabilistically derived failure projections were developed as a function of the asset management strategy being considered. Consequential costs were then factored in to obtain the business risk for each strategy. Associated with each strategy is an investment cost. In theory, if investment costs are increased the business risk should decline. At some point, carrying out analyses for several asset management strategies could identify an optimum trade-off between business risk and investment. The key process in the business case analysis was the ability to project the number of likely failures going forward, as a function of selected asset management strategies. In this study, the underlying cumulative and density distributions for the derived hazard rates were used in a Monte Carlo process to select the ages of the failed units. The process biases the random selection of failed units such that the units that are selected fit the distribution that corresponds with the derived hazard rate function. The demographic distribution is then adjusted appropriately to move ahead in time. Using this methodology, calculations were performed for each proposed strategy on a year by year basis whereby the number of units that will fail is determined for the first year, the demographic distribution is adjusted to compensate for the failed units, and then the new demographic distribution convolved with the hazard rate function to calculate the number of failures in year two. The process of projecting the number of future
5-14
failures was continued for the number of years required for the associated business case analysis. Representative results are shown in Figure 5 -3.
Projected Failures For Alternative Maintenance Strategies - Cumulative Trend
120 Number of Replacements 100 80 60 40 20 0 0 5 10 15 20 25 Years Ahead
Proactive Maintenance Strategy Minimal Maintenance Strategy Difference Betw een Strategies
Figure 5-3 Combining Demographic Data with Estimated Failure Rate Curves for Alternative Strategies
While most of the required business case analysis in the study related to decisions about a whole fleet or a large group of assets, it should be noted that the hazard rate function could also be used in business cases related to single assets. If for example there is an interest in a decision around a specific transformer, and it is known that this transformer is a member of a population for which a representative hazard rate function is known, then the hazard rate function can be used to project the probability of failure for that transformer in any given year in the future. A typical case might relate to the assessment of the business risk associated with deferring investment in a transformer. In this case the business risk per year is the sum of the expected consequential costs of failure, the expected capital cost and the maintenance cost per year. The expected consequential cost is the product of the consequential cost if failure occurs times the probability that a failure occurs. Similarly the expected capital cost is the product of the probability of failure times the inflated capital cost in the year that failure might occur. The probabilities of failure may be obtained for the calculation from the hazard rate functions for transformers of interest as described above. The primary objective for this project was to develop a methodology to address the problems presented by aging transformer populations with significant numbers of units nearing the end of their service life. The scope also included investigating the following: Longer-term view on maximizing return on operating and capital investment Relate the effect of available options on projected failure and replacement rates and their associated costs and impacts Identify emerging and hidden problems with specific segments of their transformer fleet
5-15
The project has identified and formulated a number of useful methodologies which have potential benefits for both tactical asset management and maintenance planning including: Analysis of candidate asset management strategies using quantitative risk-based approaches that can reduce costs and improve other corporate performance measures Improved replacement needs projections that can facilitate savings through improved strategic procurement arrangements and spares assignment Improved credibility of capital requirement projections for senior management and regulatory scrutiny
The project demonstrated that application of advanced decision-making methodologies is practical for some studies in terms of the quality and extent of data required, and feasible in terms of the formulation of appropriate methodologies. The underlying probabilistic and financial techniques used in this work are well founded and primarily generic. The primary challenge is to establish acceptably accurate hazard rate curves for the equipment populations of interest. Lastly, several innovative methodologies were developed and formulated for extension to the types of scenarios and business case studies that typical utility asset and maintenance managers and planners need to analyze. These include: Improved spares projection based on transformer fleet condition Forward looking failure rate projection More versus less proactive transformer replacement Tradeoffs in rating and loading capability versus maintenance and Analysis of solution options for generic transformer failures.
Although this study was limited in scope, focusing on one particular transformer type and application, it clearly demonstrated the power of combining equipment expertise and economic analysis. The project sets the stage for further integration of equipment asset management with PFM. A fundamental principal of PDAM is that investments can be fully justified only where benefits clearly outweigh costs. Aging assets is no exception and methodologies such as those described above are required to provide the basis for planning maintenance and replacement strategies. Equipment Performance Data Base The discussion above showed the usefulness and importance of having accurate equipment hazard rate curves. For the example presented, the utility had sufficient historical data for the transformer type studied but that is often not the case. Another EPRI effort is underway to gather the data needed for the methodology discussed and other PDAM and PFM applications. 8 This project, Equipment Performance Database for Transformers (IDB) can help provide important information to support the planning process.
5-16
The previous section illustrated how effective planning, maintenance, refurbishment, and replacement decisions require knowledge about asset performance and the ability to project future performance and apply the concepts of risk management to power transformers. Risk management expresses the relationship between the consequence of failure to all stakeholders and their acceptance of a failure in terms of probability of occurrence. The translation from accepted risk into performance and maintenance requirements is of great importance to power delivery organizations for both planning and operations. To be well informed about a power transformers expected performance, one must analyze both the assets individual historical performance data and that of other assets of similar characteristics or type. Similarly, as described in the preceding section, fleet management decisions are best made with an understanding of expected performance of the group. Transformer performance varies considerably because of differences in design, manufacturing, and application. Therefore, risk identification using generic transformer failure rates is not sufficient to meet the current business and technical demands for risk management placed on power delivery planners, asset and maintenance managers. Required is statistically valid information identified by: Failure type Operational history Maintenance history Design Based on specific: Family Make Model Application Age
To this end, EPRI has undertaken work to collect the needed data. By including data from numerous utilities, meaningful reliability, maintenance, and operating statistics can be generated. Some of the uses include: Population age distribution Failure mode distribution Lifetime distributions or probability density function (hazard rates)
EPRIs Transformer IDB is a collaborative effort to pool appropriate transformer operating and failure data in order to assemble a statistically valid population of many types of transformers. In a generic sense, the large component populations of distribution systems should lend themselves to statistical performance analyses with the proper groupings. With sufficient data, it will be possible to develop hazard rate curves for various PDAM and PFM applications, including that described above.
5-17
Chapter References 1. IEEE, IEEE 493-1997: IEEE Recommended Practice for the Design of Reliable Industrial and Commercial Power Systems. New York: Institute of Electrical and Electronics Engineers, 1997.], 2. G.J. Anders and J. Endrenyi C. Yung, Risk-based Planner for Asset Management, IEEE Computer Applications in Power October 2001 Volume: 14, Issue: 4 On page(s): 20-26 3. G. Anders, M. Vainberg, and M. Lebow, Extending equipment life: an application of asset management planner, Cigr Symposium on Working Plant and Systems Harder, London, UK, June 1999 4. Determination of the Actual Condition of the Electrical Components in Distribution Systems, Xiang Zhang et al Proceedings of the XIVth International Symposium on High Voltage Engineering, Tsinghua University,Beijing,China,August 25-29,2005) 5. Richard E. Brown, George Frimpong, and H. Lee Willis, Empirical Life Models, IEEE Transactions On Power Systems, VOL. 19, NO. 2, May 2004 6. Decision Support Methodology for Circuit Breaker Replacement: Part 1 Tech UpdateInformal Report. EPRI, Palo Alto, CA: 2006. 1012389. 7. Innovative Methods for Managing Aged Transformer Fleets, EPRI, Palo Alto, CA. 2005. 1011437 8. Equipment Performance Database with Common Information Model (CIM) Data Models and Performance Data for Transformers, EPRI, Palo Alto, CA: 2005. 1010592
5-18
6
CONCLUSIONS AND RECOMMENDATIONS FOR ADDITIONAL WORK
Conclusions Previous work (Technical Update: Performance Focused Maintenance - 2005, EPRI, Palo Alto, CA: 2004. 1008673.) introduced Performance Focused Maintenance (PFM) as a new approach to maintenance to meet the continuing competitive challenge of improving maintenance performance while reducing costs. The objective of the work reported upon here is to further develop the PFM concepts and integrate them into the larger set of asset management principles that are guiding power delivery organizations today. Traditionally, maintenance has been based on preserving the physical assets and striving to achieve the highest possible equipment reliability. This often emphasizes preventive maintenance schedules that contain generic tasks and frequencies and which may lead to ineffective maintenance either through inadequate or unnecessary work. PFM would suggest achieving a level of maintenance that matches the strategic service level requirements by understanding the total performance contributions of the equipment and the maintenance program to higher-level objectives. The goal of PFM is to help maintenance and asset managers direct their limited resources to the maintenance tasks that best contribute to reaching the organizations business goals. Three objectives were established and achieved in this second phase of the evolution of PFM. Merge the concepts of PFM with the guiding principles and precepts of power delivery asset management. Expand on the definitions and measures of performance in relation to PFM. Illustrate the use of PFM concepts through application to distribution substation maintenance.
Finally, this work proposed steps to continue development and enhancement of the PFM concepts.
Recommendations Additional work is required to fully realize the potential of Performance Focused Maintenance. Data PFM relies heavily on the availability of sufficient and accurate data. In practice information from a number of different databases would be required to fully implement PFM. Utilities differ greatly in the types of data stored. Additional variations can be expected in many other areas including:
6-1
Storage medium (e.g. paper records) Data precision Equipment covered Retention length
Projects of interest in this area are briefly described below. Several case studies could clarify what data can be expected to be generally available and provide examples of how a utility can combine and mine existing databases for PFM applications. Best practices for internal tracking of equipment failures have not been established. The minimum data set required for analyzing aging could be developed and documented. Such a database would facilitate not only the determination of component reliability indices, but also the analysis of the influences of component age and maintenance history. This information also would support model development listed below. Developing the ability to collect accurate performance data on substation equipment will lead to more effective maintenance decisions. Extend the current EPRI efforts to develop an Industry-wide Equipment Performance Database from the initial focus of transmission equipment to distribution equipment. This work would complement the above effort.
Equipment Models The need for better knowledge about equipment deterioration and corresponding models, especially in distribution systems, goes well beyond PFM. Ideally, such models should provide estimates on equipment failure probabilities based on input data such as equipment age, operational stresses and maintenance activities and histories. These models are needed to make the best maintenance decisions and to address the aging asset issue most efficiently and effectively. Starting with the three promising efforts discussed in the report body, continue the development of equipment models that use readily available information and provide actionable indications for asset management decisions. The goal would be to predict the equipment reliability resulting from different asset management and maintenance options.
Pilot Studies The basic concepts of PFM have been developed and documented. There is little practical value in further hypothetical studies. The great variations in power delivery organizations performance requirements, resource constraints and equipment needs suggest that pilot studies with real-world problems offer the best prospects of moving forward. In partnership with sponsoring utilities, select several case studies that address known problem areas and apply PFM solutions. Document the effort required and results achieved.
6-2
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. TogetherShaping the Future of Electricity
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 101012442
Electric Power Research Institute 3420 Hillview Avenue, Palo Alto, California 94304-1338 PO Box 10412, Palo Alto, California 94303-0813 USA 800.313.3774 650.855.2121 [email protected] www.epri.com