Current practices and customer value-based distribution system reliability planning

1174 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 40, NO. 5, SEPTEMBER/OCTOBER 2004

Current Practices and Customer Value-BasedDistribution System Reliability Planning

Ali. A. Chowdhury, Senior Member, IEEE, and Don O. Koval, Fellow, IEEE

Abstract—Among the major issues facing utilities in today’scompetitive electricity market is the pressure to hold the lineon rates and provide electricity with adequate quality and re-liability. Utilities are increasingly recognizing that the level ofsupply reliability planned and designed into a system has toevolve away from levels determined basically on a technicalframework using deterministic criteria, and toward a balancebetween minimizing costs and achieving a sustainable level ofcustomer complaints. Assessment of the cost of maintaining acertain level of supply reliability or making incremental changestherein must include not only the utility’s cost of providing suchreliability and the potential revenue losses during outages, but alsothe interruption costs incurred by the affected customers duringutility power outages. Such a cost–benefit analysis constitutes thefocal point of the value-based reliability planning. Value-basedreliability planning provides a rational and consistent frameworkfor answering the fundamental economic question of how muchreliability is adequate from the customer perspective and wherea utility should spend its reliability dollars to optimize efficiencyand satisfy customers’ electricity requirements at the lowestcost. Costs to customers associated with varying levels of servicereliability are significant factors that cannot be ignored. Explicitconsiderations of these customer interruption costs in developingsupply reliability targets and in evaluating alternate proposalsfor network upgrade, maintenance, and system design must,therefore, be included in system planning and design process. Thepaper provides a brief overview of current deterministic planningpractices in utility distribution system planning, and introducesa probabilistic customer value-based approach to alternate feedrequirements planning for overhead distribution networks.

Index Terms—Configurations, cost–benefit tradeoff, customerinterruption cost, deregulated competitive market, deterministicreliability targets, failure modes, industrial, interruptions, opencircuit, reliability, short circuit, substations, switching, value-based distribution planning.

I. INTRODUCTION

THE determination of acceptable levels of supply reliabilityis presently achieved by comparing actual interruption fre-

quency and duration indexes with arbitrary targets [1]–[6]. In-herent in this approach is the perception of customer satisfaction

Paper ICPSD-IAS-82-4, presented at the 2000 Industry Applications SocietyAnnual Meeting, Rome, Italy, October 8–12, and approved for publication inthe IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS by the Power SystemsEngineering Committee of the IEEE Industry Applications Society. Manuscriptsubmitted for review October 15, 2000 and released for publication June 4, 2004.

A. A. Chowdhury is with Transmission and Distribution Planning,DAV5, MidAmerican Energy Company, Davenport, IA 52801 USA (e-mail:[email protected]).

D. O. Koval is with the Department of Electrical and Computer Engi-neering, University of Alberta, Sherwood Park, AB T8A 2G9, Canada (e-mail:[email protected]).

Digital Object Identifier 10.1109/TIA.2004.834075

level for supply cessation. This type of implicit reliability cri-teria is inadequate for rationally evaluating the validity of sug-gested capital investment to materialize improvements that opti-mize utility efficiency. It is, therefore, a foregone conclusion thatrules of thumb and vague criteria cannot be applied in a consis-tent manner to the very large amount of capital investment andoperating and maintenance decisions that electric utilities rou-tinely make. The ultimate impact is a likely misallocation of re-sources within distribution systems.

In order to provide a rational and consistent means of pru-dent decision making on the necessity of changing supply re-liability levels experienced by utility customers in a given ser-vice area, quantifiable factors other than utility revenue lossesrequired to be modeled. In particular, explicit modeling of cus-tomer damage costs in establishing supply reliability criteria hasto be incorporated in the regular planning practices.

A value based reliability planning methodology attempts toascertain the minimum cost solution, where costs are identifiedas the sum of investment cost plus operating and maintenancecost plus customer outage costs. This minimum cost point isnormally defined by the marginality condition where the mar-ginal cost of reliability enhancement is equal to the marginalbenefit, which is the expected reduction in customer damagecosts due to the marginal investment. This paper presentsa brief overview of current utility practices in distributionplanning. In particular, deterministic planning criteria used inalternate feed requirements planning and design for overheadnetworks are detailed. A probabilistic customer value addedalternate feed planning approach that could complement thecurrent deterministic criteria is illustrated with a numericalexample system.

II. CURRENT DISTRIBUTION PLANNING AND DESIGN CRITERIA

The primary criteria for distribution system planning and de-sign used by most utilities are loading and voltage [4]. It is im-portant to note that utilities have traditionally considered relia-bility qualitatively in some fashion in their system planning anddesign. Losses and environment also are considered by someutilities as a secondary criterion. The planning criteria wereestablished by utilities, based on a host of considerations in-cluding industry practices, electrical and construction codes,manufacturers’ equipment ratings, rules of thumb developedwith long term operating experience. It is worth noting that atpresent, though limited, some utilities have started utilizing cus-tomer value-based distribution reliability methodologies in reg-ular planning activities for project justification.

0093-9994/04$20.00 © 2004 IEEE

CHOWDHURY AND KOVAL: CURRENT PRACTICES AND CUSTOMER VALUE-BASED DISTRIBUTION SYSTEM RELIABILITY PLANNING 1175

A. Outage Data Collection and Reporting

Virtually all North American utilities have some form of acomputerized database and interruption reporting system to logkey data elements on component and feeder failures and cus-tomer outages. Customer outages are identified by time of oc-currence, duration, weather condition, and by causes that can begenerally categorized as planned, equipment failure, trees, for-eign objects, human error, lightning, supply, etc.

Operating experience records from the field are normally en-tered into a database and a number of reports are generated.These reports serve basically three purposes in a distributionfunction: reporting, planning, and maintenance such as vege-tation control or tree trimming.

B. Reliability Indexes

The most commonly used indexes of distribution system re-liability are one or more of the following system and customerindexes [4]–[6]:

System Average Interruption Frequency Index (SAIFI)

SAIFItotal number of customer interruptions

total number of customers served

SAIFI is the average number of interruptions per customerserved.System Average Interruption Duration Index (SAIDI)

SAIDIsum of customer interruption durations

total number of customers served

SAIDI is the average duration of a customer interruption,per customer served.Customer Average Interruption Duration Index (CAIDI)

CAIDIsum of customer interruption durationstotal number of interrupted customers

CAIDI measures the average duration of a customer inter-ruption within the class of customers that experienced atleast one sustained interruption.Customer Average Interruption Frequency Index (CAIFI)

CAIFItotal number of customer interruptionstotal number of interrupted customers

CAIFI defines the conditional average number of interrup-tions among the class of customers who experience at leastone interruption.Average System Availability Index (ASAI) is the ratio oftotal customer hours that service was available, divided bythe total customer hours in the time period for which theindex is calculated. On an annual basis, it can be shownthat

ASAI 1SAIDI

8760 hours per year100

To illustrate, an ASAI of 99.9543 indicates that the averagecustomer had service available in 8756.0 h out of a total8760 h in the year, i. e., SAIDI h/year.

These indexes are routinely reported on an annual basisand are typically computed for each feeder. Aggregate reportspresent comparable statistics on a district or regional basis andon a system-wide basis as well. These indexes can be readilycomputed down to the distribution transformer level.

C. Targets for Customer Service Reliability

The historical reliability indexes are used to establish guide-lines by many utilities in planning and design. The reliabilityof service is measured by SAIFI, CAIDI, SAIDI, and ASAI inmost cases [5]. Some utilities use different target levels for theoverall system and for the districts or regions. It is, however,worth noting that none of the utilities have minimum standardsfor customer service reliability. Most utilities prefer to have ob-jectives or targets they strive toward. The basic distinction isthat a standard is perceived to be a “hard constraint,” whereas“guidelines,” “objectives,” and “targets” refer to performancelevels that are considered to be desirable and that a utility strivestoward [4]–[6]. Typical numerical values of the reliability in-dexes used as desired targets by different utilities are summa-rized in Table I. A typical district and company service pointtargets are shown in Table II

As shown in Table I, some utilities formally use quantitativetargets for SAIFI, SAIDI, CAIDI, or ASAI. The rationale forthe target values is that these figures are consistent with actualperformance during a certain historical period when service reli-ability to customers was considered to be adequate by the utilityand the associated level of customer complaints was viewed asnot being excessive. Two examples of mandated distribution re-liability standards by regulatory bodies in a deregulated envi-ronment are presented in Section II-D.

D. Examples of Distribution Reliability Standards in aDeregulated Market

Customers are increasingly looking for improved service re-liability from electricity suppliers. This has been recognized byregulators who incorporate customer reliability measures suchas “how often” and “how long” customers can be interruptedin year. For example, as a part of the deregulation of electricutilities in California, the State established in 1996 a new ratemechanism, called “Performance Based Ratemaking” (PBR).To guarantee reliable service to the customers PBR includesperformance incentives with a specific system of rewards orpenalties for each utility. With respect to Southern CaliforniaEdison Company, the PBR rewards or penalties are based onthe two-year rolling averages of the annual customer minutes ofinterruptions (CMI or SAIDI) and the annual number of distri-bution circuit interruptions. Catastrophic events are excluded inboth CMI and distribution circuit interruptions computations,however, storms events are included. The PBR guidelines in-clude three ranges, upper, middle and lower for the CMI anddistribution circuit interruptions (DCI). The CMI and DCI fig-ures for Southern California Edison are shown in Tables III andIV [7].


TABLE ITYPICAL DISTRIBUTION RELIABILITY TARGETS USED BY UTILITIES

TABLE IIDISTRICT AND COMPANY SERVICE POINT TARGETS

These limits became effective in 1997 and are being reducedby two CMI for the next five years.

In the Australian State of Victoria, the regulatory bodies havewritten into the Distribution Code the following performancerequirements [8], [9].

1) “A local distribution company must use reasonable en-deavors to ensure that the duration of interruption of the

TABLE IIISOUTHERN CALIFORNIA EDISON CMI INDEXES

TABLE IVSOUTHERN CALIFORNIA EDISON NUMBER OF DCI INDEXES

supply of electricity to a customer’s installation does notexceed on average 500 min (8.3 h) per annum in ruralareas, and 250 min (4.15 h) per annum in other areas.”


2) “On request a local distribution company must make indi-vidual customer targets and actual performance informa-tion available to the customer.”

It is important to note that these deregulated criteria as wellas conventional targets presented in Table I, in no instance, arelinked to or derived from estimates of the value of service relia-bility of customers [4]. The value of service information has notbeen used explicitly to rank order the reliability performance ofdifferent feeders and identify poor performance for upgradingbased on total interruption costs incurred by customers on thefeeder.

III. RELIABILITY COST VERSUS RELIABILITY BENEFIT

TRADEOFFS IN DISTRIBUTION SYSTEM PLANNING

Utilities routinely make many reliability-related investmentand operating decisions. This section presents typical cost–benefit tradeoff situations in distribution system reliabilityplanning, where the utilization of customer interruption costdata can enhance decision making to a more rational, con-sistent, and economic framework. The distribution systemconsists of a very large number of individual components (e. g.,subtransmission circuits, distribution stations, primary feeders,distribution transformers, secondary circuits and customers’connections). As a result, a large number of standards suchas design, construction, operation, maintenance, etc., havebeen developed over the years that entirely define supplyreliability provided to utility customers. It is important to notethat the current different distribution planning practices canbe expanded to a value-based planning framework. A hostof distribution planning and operating practices can benefitfrom a value-based planning approach. For example, there aremany distribution system reliability problems, where customerinterruption cost information could be explicitly utilized tosupport rational investment decision making, such as, facilitiesdesign and configuration, feeder upgrade, distribution systemupgrading, tree trimming, pole maintenance, protection design,service restoration, underground cables design, cable replace-ment, distribution system automation, general system-wideutilization, etc. [4].

As a result of load growth, completely new distribution fa-cilities are required in certain instances. This may be the casein surrounding areas of large cities. In these situations, deci-sions have to be made regarding substation configuration, sizeand number of transformers, capacity planning, selecting routefor a new line, etc. Regarding route selection, if two routes areavailable for an overhead line extension, and one route is morecostly but has less exposure to traffic, decision has to be madewhether the more expensive route is justified from customers’benefit point of view.

Utilities normally identify poorly performing feeders as po-tential candidates for feeder upgrading using one or more im-plicit criteria. These may include the number of complaints re-ceived or other deterministic decision rules based on deviationsfrom calculated SAIFI, SAIDI, etc. Normally, the number offeeders identified for upgrading and the investment made forthis purpose is further constrained by capital and operation andmaintenance budget, and manpower resources.

Feeder upgrade decisions made in the context of poorly per-forming feeders could be made on a more rational and consis-tent manner even within severe budget constrained situations byrank ordering the feeders on the framework of the total costs ofimprovements relative to the benefits. The benefit would be as-sessed by quantifying expected reductions in interruption coststo all customers as a consequence of the upgrade being con-sidered. Using the cost–benefit strategy, the available resourcescould be more objectively and consistently utilized across thefeeders since such a value-based method facilitate establishingan economically rational prioritization of the feeders requiringupgrades. It is important to note that in value-based approach,there are no absolute standards of SAIFI, SAIDI, CAIDI, etc.,that could be utilized uniformly across the feeders. In a value-based planning framework, rather, customers’ preferences andmix unique to each feeder and/or geographical service area de-termine an inadequate or poor performance. In competitive elec-tricity market, value-based approach, therefore, makes perfectsense.

With load growth and customer mix change, the distributionsystem requires upgrading. One aspect of such changes is thecurrent trend toward a higher distribution voltage level wherepossible. Conductor upgrading, installation of line regulatorsto boost voltages when required, transformer upgrading to ac-commodate load increases, replacement of under-loaded trans-formers with smaller transformers and substation loading cri-teria (e. g., high loading can result in extremely poor reliability)are a few of the typical decisions related to system upgrading.

Vegetation control normally affects the service reliabilitylevel of a distribution system. The amount of work done oneach occasion, and the frequency of maintenance are the twomajor decisions have to be made in regard to tree trimming.Reliability centered tree trimming could result in optimumtree trimming schedules balancing the marginal cost and themarginal benefit from customers’ perspective.

Normally the most lines in a distribution system are overheadwires on wooden poles. The poles require continuous mainte-nance. The level of maintenance for poles and their replacementprograms affect the level of distribution system reliability.

An increase in protection, such as number of devices, typeof devices, location of devices will inevitably result in changesin distribution system supply reliability. Increases in protectionspending will protect the distribution system from unreliabilityresulting from sabotage, lightning, and equipment failure bylimiting the area impacted by these events and by limiting theirfrequency and duration. In turn, this will result in lower totaloutage cost to customers. Examples of investment decisions re-lated to protection include; fuse coordination, installation of pri-mary fuses, installation of switches to facilitate sectionalizing,dual-feed substations versus single feed, fault indicators, etc.

From customers’ point of view, supply reliability means howquickly voltage is restored following an interruption. It couldbe different for different customer types. Strategies related tothe sizing, location and scheduling of service restoration crewsas well as service restoration sequence by feeder could have animportant impact on the outage costs incurred by customers.

Most distribution system lines are overhead lines requiringregular tree trimming and pole maintenance. The reliability


trade-offs in the overhead versus under-grounding involve alower frequency of customer interruptions in the undergroundcase. Customer outages due to underground equipment failures,however, can be longer in duration due to the time required tolocate and rectify the problem.

Virtually all utilities use rules of thumb in their cable replace-ment strategies. Typical considerations for cable replacementare, replace after a certain number of failures, number of cus-tomers and cost to replace, moisture content, number of failures,and number of customers served, length, age, commercial or in-dustrial customers, etc.

The electrical energy to customers through distributionfeeders could be more cost-effectively and reliably delivered byautomation of the feeders. Generally most feeder interruptingdevices are static and requires line crews to open and closeswitches manually in order to locate and isolate faults, andrestore voltage to customers. In addition, as the distributionnetwork becomes more and more complex due to expansion,conventional feeder maintenance practice becomes inadequateand customers have to endure longer outages. Because of thetime required in locating faults manually, revenue losses tothe utility could also be significantly high. Extended durationoutages have serious monetary impacts from the customer’sperspective as well. Distribution automation can provide acost-effective means for reducing the frequency and durationof potential service interruptions.

A number of typical decisions where customer interruptioncosts and value-based methodology could be utilized are as fol-lows: 1) for assessing the appropriate level of reliability forfeeders in each district/region. This should be based on the ex-pectations of the customer mix on the feeder/region and con-sistent with their willingness to pay; 2) for assessing the levelof distribution reliability that customers are willing to pay for;this means, analyze the distribution system as whole and com-pute the total dollars required to raise or lower the reliability ofthe distribution system to the level customers are willing to payfor; and 3) for determining the amount of corporate reliabilitydollars to be spent on distribution versus other segments of thepower system.

Under current planning and design practices, investmentdecisions in the context of any of the number of reliabilityrelated problems identified earlier are based on implicit rulesof thumb criteria normally employed by utilities. Explicitcost–benefit analysis is not always undertaken. Though, forthe purpose of rank ordering problem feeders, avoided lostrevenues to the utility provide a measure of the benefits ofupgrading, present planning and operating practices, as theyrelate to distribution service reliability, are not based on a directand objectively specified linkage between the level of reliabilitythat is planned and delivered and the level of reliability thatcustomers want. This section presents a typical utility’s currentpractices in alternate feed requirements planning for overheaddistribution systems. An illustrative customer cost–customerbenefit assessment for justification of an alternate feed to amajor load center is performed to demonstrate the underlyingconcepts of reliability cost-reliability worth analysis that couldcomplement the current deterministic planning practices indistribution investment decision makings.

TABLE VPEAK LOAD RANGE VERSUS PERCENT OF COINCIDENT PEAK

DEMAND GUIDELINES

IV. ALTERNATE FEED REQUIREMENTS FOR OVERHEAD

DISTRIBUTION SYSTEMS

The standard service provided to all overhead customers is ra-dial feed. Full or partial alternate is normally considered whenthe additional cost is paid by the customer. A rudimentary evalu-ation of the advantages to the customers and the utility is lookedat and a comparison of the costs of providing alternate feedis made [4]–[6]. Evaluation of costs normally includes the ad-ditional costs of the transmission, substation, and breakers re-quired. Partial alternate supply having 80% capacity on peak isdeemed as fulfilling the requirements. Generally, on the distri-bution system communities with a certain demand level such as3000 kVA and higher is considered for alternate feeds. Utilitypersonnel look at acceptable restoration times for different typesof customers under various outage conditions. Situations whereoutage times exceed the acceptable restoration times are exam-ined and a contingency plan developed to deal with the spe-cific situation. For communities with certain load demand (e. g.,3000 kVA and higher) this could mean an alternate feed. Cus-tomers fed from underground residential subdivisions, by natureof the design and development of the system, are usually pro-vided with an alternate feed.

The existing criteria for determining the provision of al-ternate feed are primarily based on deterministic approach[4]–[6]. As stated earlier, deterministic approaches are notbased on a formal framework, rather based on planner’s ex-perience and intuition which do not and cannot account forthe probabilistic nature of distribution system behavior, of cus-tomer demands or of system component failures. Since thebasic objective of distribution planning criteria is to provide aconsistent approach to obtaining a balance between the distri-bution system performance and the total cost to satisfy bothcustomer and a utility need, alternate feed requirements iden-tified solely based on deterministic criteria have to be verifiedthrough probabilistic value-based analyses that include cus-tomer interruption costs.

This involves the recognition of reliability cost/reliabilitybenefit. A cost/benefit analysis to determine when an alternatedistribution feed should be planned is needed to be performedto complement the deterministic guidelines for distributionplanning. The alternate source would likely be provided if theanalysis indicates that the improvement in supply reliabilitywould be cost effective. Ideally, outage probability and outagecost data specific to a utility distribution system required tobe utilized for cost/benefit analysis. In this paper, an example


TABLE VIPERCENTAGE OF TOWN LOAD THAT SHOULD HAVE A BACKUP SUPPLY VERSUS TOWN POPULATION

Fig. 1. Illustration of guideline A.1.1.

of using customer interruption cost data in evaluating the eco-nomics of alternate distribution feed is described. It is importantto note that this analysis includes only forced outages. Mainte-nance outages are not modeled in the analysis because of thefact that impacts of maintenance outages can be controlled andminimized through advance communications of maintenanceschedules to customers, scheduling maintenance outages duringlight load periods, performing hot line maintenance, bringing inmobile subs, maintenance scheduling during customer’s downtimes, etc.

A. Examples of Deterministic Planning Guidelines forAlternate Feed Requirements

1) Reliability of Supply to 25-kV Buses: The distributionsupply system whose peak load is less than 10 MW is plannedwith backup capacity to a minimum of 30% of coincident peakdemand as shown in Tables V and VI by means of transformerstandby or through the ability to import standby capacity fromadjacent substations or a combination of both of these. It is an-ticipated that a failed transformer should normally be replacedwithin 12 h. This guideline is based on the assumption thatthe 20% of the feeder peak demand would cover all essentialservices for different customer types. A typical 25-distributionnetwork is depicted in Fig. 1. According to this guideline, inthe event of a fault on 25-kV line “A” or in the upstream of

it, 25 kV line “B” while operating at its peak demand, wouldbe capable of providing supply to 20% of the peak demand of25 kV line “A.” During contingency, voltage will be allowed todrop to extreme values (e.g., 0.90 pu).

Whenever possible the distribution system is arranged to ac-commodate rotational supply to those customers who are inter-rupted.

2) Reliability of Supply to Towns/Cities: The level of backupsupply to a town or a city suggested in the following reflects thesize of the town or the city. The underlying assumption is that asa population center gets larger, a greater number of customersget impacted due to cessation of power supply, and the cost percustomer to alternate supply generally becomes lower.

3) Reliability of Supply to Large Users and Industrial Cus-tomers: The normal supply to large users and industrial cus-tomers is radial feed. An alternate feed will be provided to thesecustomer types at full customer costs.

B. Value-Based Alternate Feeder Requirements Planning

Deterministic criteria given in Section IV-A alone are insuf-ficient for rationally assessing the validity of suggested alter-nate feed requirements. There is an increasing recognition in theelectric utility industry that investments related to the provisionof electric service reliability should be more explicitly evaluatedconsidering their cost and benefit implications. The underlyingintention is to relate the benefit of uninterrupted power supply asa means to rationalize the cost of alternate feed additions. Sucha cost benefit assessment is the focal point of the probabilisticvalue-based reliability planning [10], [11]. As mentioned ear-lier, a value-based reliability planning approach attempts to lo-cate the minimum cost solution where the total cost includes theutility investment cost plus the operating and maintenance costplus the customer interruption costs. The underlying principleof value-based planning is illustrated in Fig. 2.

Fig. 2 illustrates how utility costs reflected in customer ratesand customer interruption costs are combined to give total cus-tomer cost. The utility cost curve shows how customer ratesgo up as more money is spent for increased distribution system


Fig. 2. Reliability cost/reliability worth.

reliability levels. The customer interruption cost curve showshow customer cost of interruptions decreases as the distribu-tion system reliability increases. It is also important to notethat for low levels of distribution system reliability levels, thecustomer interruption costs are significant. However, the utilitycost can also increase significantly in the additional costs ofrestoring the system to a normal operating state and the lossof revenue (i. e., the utility cost curve shown in Fig. 2 isbased on the belief that increased costs will achieve higherlevels of distribution system reliability). When the combinedutility and customer interruption costs are minimized, then theutility customers will receive the least cost service. Therefore,using the concept of value based distribution system reliabilityplanning, a given level of service reliability can be examinedin terms of the costs and the worth to the customer of pro-viding the electric service from various proposed distributionoperating configurations.

The basic assumption in value based assessment is thatplanning practices as they relate to distribution service reli-ability, should be based on a direct and objectively specifiedrelationship between the level of reliability that is planned anddelivered, and the level of reliability that customers expect. Anyinvestment decision related to distribution reliability, however,could easily be addressed on a more rational and quantitativebasis within a cost–benefit framework explicitly using data onthe value of changes in service reliability to customers [10].The ultimate determination of poor or adequate distributionreliability performance in value-based planning framework isbased on customer’s preferences and customer mix that areunique to each feeder. In a competitive market environment,such an approach makes perfect sense.

1) Customer Interruption Cost Data: The value of service,i.e., the worth of reliability expressed in terms of costs of cus-tomer interruptions can be established based on actual surveysof customer perception of regarding the level of service relia-bility they are willing to pay for. By establishing a method ofgiving a dollar value to various levels of service reliability, itis possible to ascertain the balance where distribution systemreliability is best matched. The data compiled from customersurveys lead to the creation of sector damage functions [11].The cost of interruptions at a single customer load point isdependent entirely on the cost characteristics of that customer.The sector damage function presents the sector interruption

Fig. 3. Utility sector customer damage functions.

costs as a function of the duration of service interruptions.The customer costs associated with an interruption at any loadpoint in the distribution system involves the combination ofcosts associated with all customer types affected by the distri-bution system outage. This combination leads to the generationof a composite customer damage function. References 10 and11 illustrate the approach involved in creating a compositecustomer damage function. The customer damage functions[12], after escalation to $2000 (Canadian) for each sector areshown in Fig. 3. The cost of interruptions is expressed in dol-lars per kilowatt.

In this paper, one simple illustrative example showing thepractical applications of the reliability cost/reliability benefitmethodology is described. A reliability cost/reliability benefitassessment to determine when an alternate feed to a major cityshould be added is required to be performed to complementthe deterministic distribution reliability planning criteria. Thealternate feed should be provided if the analysis indicates thatthe improvement in service reliability would be cost effective.In practice, outage probability and interruption cost data spe-cific to the distribution network of study or the best represen-tative data available at the time of analysis should be utilizedfor the reliability cost/reliability worth analysis. An actual ap-plication should consider the amount of load growth and shouldshow numbers on an escalated and present worth basis for fu-ture years.

2) An Illustrative Example for Justification of an AlternateFeed to a Major City: A numerical example illustrating the ap-plication of reliability cost/reliability benefit approach in evalu-ating the cost effectiveness of an alternate feed to a major cityis presented. A very simplified representation of the investmentdecision is adopted in this example. The example uses genericoutage and cost information data. The basic objective of this ex-ample is to focus on the framework for cost–benefit analysis insimilar situations.


TABLE VIIAN ALTERNATE FEED TO BE ADDED TO A MAJOR CITY

For this example, assume that there is a major city with a pop-ulation of over 25 000. Deterministic Planning Criteria underSection IV state that 80% of this city load should have a backupsupply. Table VII describes the steps to calculate the presentvalue of reliability benefits, i.e., avoided customer interruptioncosts, should an alternate feed be provided to this load center.

Table VII is self-explanatory. It is assumed that the alternatefeed’s economic life is ten years. The city’s current unreliabilityshould be based on historical experience data that is assumed tobe 5.0 h per year in this example. Even with the installation ofan alternate feed, it will take 1 h to switch power supply fromthe alternate source. Therefore, the expected number of outagehours avoided is h per year.

If the unserved load is 10 MW with a load factor of 85%,then the avoided expected unserved energy is 27.20 MWh con-sidering only 80% of the city load would have a backup supply.Assuming the customer interruption cost for the mix of cus-tomers served by the city to be $14 per kWh, yields a presentvalue of reliability benefit of $2 882 656 at 0% load growth, and$3 122 560 at a 2% load growth. The present value figures werecalculated assuming 10% discount rate and 3% inflation rate.These benefits together with loss reduction benefits if any, canbe compared to the present value revenue requirements to installthe alternate feed in question.

V. CONCLUSION

Current utility practices in distribution system reliabilityplanning are presented in this paper. It is important to note thatneither traditional planning criteria nor the current mandatedcriteria by regulators in a deregulated market are linked tocustomer preferences.

Typical utility distribution planning practices in determiningalternate feed requirements for overhead distribution systemsare presented. Deterministic planning practices are comple-mented with a probabilistic value based planning analysis inthe cost effective assessment of an alternate feed to a majorload center.

In a competitive energy market in which reliability of servicedoes influence customer purchasing decisions, a utility can notafford to ignore customer preferences. Today’s energy marketis characterized by intense price competition under continuouspressures on utilities to hold the line on rate increase. Valuebased planning renders a rational solution to these emergingpressures and will permit service reliability to evolve toward alevel that customers would perceive to be a fair value.

REFERENCES

[1] M. Munasinghe and M. Gellerson, “Economic criteria for optimizingpower system reliability levels,” Bell J. Econ., pp. 353–365, Spring 1979.

[2] “Cost-benefit analysis of power system reliability: Determination of in-terruption costs—Vol. 1: Measurement methods and potential applica-tions in reliability cost–benefit analysis,” Elect. Power Res. Inst., PaloAlto, CA, EPRI IL-6791, 1990.

[3] M. J. Sullivan, B. N. Suddeth, T. Vardell, and A. Vojdani, “Interrup-tion costs, customer satisfaction and expectations for service reliability,”IEEE Trans. Power Syst., vol. 11, pp. 989–995, May 1996.

[4] “Economic Penalties Including Customer Costs for Loss of Service Con-tinuity,” Canadian Elect. Assoc., Montreal, QC, Canada, Res. Develop.Rep. SD-273, 1991.

[5] R. Billinton and J. E. Billinton, “Distribution system reliability indices,”IEEE Trans. Power Delivery, vol. 4, pp. 561–586, Jan. 1989.

[6] “1993 annual service continuity report on distribution system perfor-mance in Canadaian electric utilities,” Canadian Elect. Assoc., Montreal,QC, Canada, 1992.

[7] P. Save and T. Velarde, “Incorporation of performance based rating(PBR) incentives in probabilistic transmission planning criteria,” inProc. 5th Int. Conf. Probabilistic Methods Applied to Power Systems(PMAPS’97), Vancouver, BC, Canada, Sept. 21–25, 1997, pp. 17–24.

[8] Victorian Distribution Code, Section 19, Interruption to Supply, 19.1(a)and (b), Office of the Regular General, Victoria, Australia, 2002.

[9] P. Lyons and G. Nourbakhs, “Applied subtransmission and distributionreliability assessment of SEQEB network in Queensland,” in Proc. 1997AUPEC Conf., Sydney, Australia [CD ROM].

[10] L. Goel and R. Billinton, “A procedure for evaluating interrupted energyassessment rates in an overall electric power system,” IEEE Trans. PowerSyst., vol. 8, pp. 929–936, Aug. 1993.

[11] G. Wacker, R. K. Subramaniam, and R. Billinton, “Using cost of elec-tric service interruption surveys in the determination of a compositecustomer damage function,” in Proc. IASTED Energy, Power and En-vironmental Systems Conf., San Francisco, CA, June 4–6, 1984, Paper203-143.

[12] G. Tollefson, R. Billinton, G. Wacker, E. Chan, and J. Awaya, “A Cana-dian customer survey to assess power system reliability worth,” IEEETrans. Power Syst., vol. 9, pp. 443–450, Feb. 1994.


Ali A. Chowdhury (A’83–S’86–M’88–SM’94)received the M.Sc. degree with honors in electricalengineering from the Belarus Polytechnic Institute,Minsk, Belarus, the M.Sc. and Ph.D. degrees inelectrical engineering with specialization in powersystems reliability and security from the Universityof Saskatchewan, Saskatoon, SK, Canada, andthe M.B.A. degree from St. Ambrose University,Davenport, IA.

He has over 25 years electric utility, electricequipment manufacturing industry, consulting,

teaching, and R&D experience in power system reliability and security assess-ments, planning, and analysis. He is actively involved in the development ofprobabilistic models, criteria, and software for use in power system planning,operating, and maintenance. He has lectured on power system reliability andsecurity nationally and internationally. He has published extensively on powersystem reliability and value-based utility system reliability planning and designtopics. He is currently with MidAmerican Energy Company, Davenport, IA.

Dr. Chowdhury is a recipient of numerous national talent scheme scholar-ships, has been listed in Marquis Who’s Who in America, Who’s Who in theWorld, is a Fellow of the Institution of Electrical Engineers, U.K., a CharteredEngineer in the U.K., and a Registered Professional Engineer in the State ofTexas and in the Province of Alberta, Canada.

Don O. Koval (S’64–M’65–SM’78–F’90) is a Pro-fessor in the Department of Electrical and ComputerEngineering, University of Alberta, Edmonton, AB,Canada. He teaches classes in Reliability Engi-neering, Power Quality, Power System Analysis,and “IEEE Gold Book.” For 12 years, he was a Dis-tribution Special Studies Engineer with B.C. Hydroand Power Authority, Vancouver, BC, Canada, and,for two years, he was a Subtransmission DesignEngineer with Saskatchewan Power, Regina, SK,Canada. He serves on the Boards of Directors of

several international societies, including the International Association ofScience and Technology for Development (IASTED) and the InternationalInstitute for Advanced Studies in Systems Research and Cybernetics (ICSRIC).He has authored or coauthored more than 250 technical publications in thefields of emergency and standby power systems, power system reliability,human reliability, power system disturbances and outages, power quality, andcomputer system performance. He was the Editor of the IASTED InternationalProceedings on High Technology in the Power Industry, 1996.

Dr. Koval is a Registered Professional Engineer in the Provinces of Albertaand British Columbia, Canada, a Fellow of the American Biographical Insti-tute, and a Life Fellow of the International Biographical Centre, Cambridge,U.K. He is listed in Marquis Who’s Who in the West, Who’s Who in America,Who’s Who in the World, Personalities of the Americans, Who’s Who in Scienceand Engineering, 5000 Personalities of the World, and in the International Bio-graphical Centre’s International Leaders of Achievement, International Who’sWho of Intellectuals, and Men of Achievement. He was Co-Chairman of the1998 IEEE/IAS Industrial and Commercial Power Systems Technical Confer-ence held in Edmonton, AB, Canada. He is Chairman of IEEE Std. 493 (IEEEGold Book). He was elected as one of the six Distinguished Lecturers of theIEEE Industry Applications Society (IAS) for the period 2000–2001. He wasalso recently appointed to the rank of Distinguished Visiting Professor Recentlyand elected Fellow by the International Institute for Advanced Studies in Sys-tems Research and Cybernetics in Germany.

Documents

Current practices and customer value-based distribution system reliability planning