Smart Grid Fault Location, Isolation, and Service Restoration (FLISR) Solutions to Manage Operational and Capital Expenditures

Transcription

1 1 Smart Grd Fault Locaton, Isolaton, and Servce Restoraton (FLISR) Solutons to Manage Operatonal and Captal Expendtures Adam Gauc, P.Eng. Schneder Electrc North Amerca Abstract Wth new requrements beng placed on the dstrbuton grd, the cost to consumers of poor supply qualty and low relablty s ncreasng. Consumers are startng to demand hgher levels of supply qualty from dstrbuton network operators. Ths paper wll dscuss varous Fault Locaton, Isolaton and Servce Restoraton (FLISR) strateges wth ncreasng CAPEX nvestment potental that can be used to mprove the qualty of supply. I. INTRODUCTION Increasng Demands Requre Increased Performance In North Amerca today, one of the key ssues facng our dstrbuton networks s the urgent need to make captal nvestments to replace and upgrade the sometmes neglected, agng dstrbuton nfrastructure. Addtonally, new technologes are placng greater demands on ths already stressed nfrastructure. Some of these new and emergng technologes nclude dstrbuted generaton, renewables and plug-n electrc vehcles. A. Dstrbuted Generaton and Renewables Many governments are now mandatng a certan percentage of ther energy portfolo n ther jursdcton must come from clean sources. Followng ths requrement, the regulatory envronment n many jursdctons has adapted to allow generaton to nterface drectly wth the dstrbuton grd. Ths s causng tradtonal energy consumers to become energy producers by allowng them to nvest n generaton such as small roof top solar systems up to md-sze wnd farms, or by sellng unused generaton capacty from co-generaton plants back nto the grd. Ths s creatng a vertable energy marketplace out of the dstrbuton system that wll requre a hgh degree of relablty as producers seek to make ther return on nvestment (ROI). B. Plug-In Electrc Vehcles Wth the ever ncreasng prce of gasolne, consumers are begnnng to look for more cost-effectve means of transportaton. The ntroducton of partally or fully electrc vehcles s seen as one of these lower cost alternatves. Consumers of new electrc vehcles wll requre the supply of suffcent energy to charge ther vehcles at ther homes and offces. Census data from both Canada and the Unted States show that 72-75% of commuters drve a vehcle as ther prmary mode of transportaton to ther workplace. Sustaned outages can have the potental mpact of nterruptng a consumer s daly commute, leadng to absenteesm and greater economc dsrupton. II. MEASURING QUALITY OF SUPPLY Because the dstrbuton system s made up of many dfferent stakeholders such as producers, consumers, traders and network operators, t can often be dffcult to determne the basc level of system supply qualty and relablty. In North Amerca, ths responsblty falls to regonal energy regulators. Regulators can entce dstrbuton network operators to meet and even exceed these levels by creatng a framework for performance-based rates. The man crtera for measurng the qualty of supply are the System Average Interrupton Duraton Index (SAIDI) and the System Average Interrupton Frequency Index (SAIFI). SAIDI measures the average amount of tme that consumers are wthout servce per consumer served and s defned by the followng equaton: r N SAIDI =. (1) N t

2 2 Where N s the number of customers affected by an outage, r s the restoraton tme of an outage and N t s the total number of customers served. SAIFI measures the total number of outages experenced by consumers per consumer served and s defned by the followng equaton: N SAIFI =. (2) N t Where N s the number of customers affected by an outage and N t s the total number of customers served. Usually outages of less than 1-3 mnutes are not counted towards the SAIDI or SAIFI metrcs. III. THE COSTS OF SUPPLY QUALITY The cost of supply qualty can be borne by both consumers of energy and dstrbuton network operators. Fg. 1. Costs versus Qualty of Supply. When consderng the nterrupton costs for consumers because of low supply qualty together wth the costs for dstrbuton network operators to mplement supply qualty mprovements, the optmum supply qualty can be found as the mnmum cost. A. Energy Consumers The cost of an outage for commercal and ndustral consumers can have a drect economc mpact on ther operatons. TABLE I TYPICAL FINANCIAL LOSS PER OUTAGE Industry Cost Semconductor Producton $5,300,000 Fnancal Tradng $8,400,000 per hour Computer Center $1,000,000 Telecommuncatons $45,000 per mnute Steel Works $500,000 Glass Industry $ Source: The Costs of Poor Power Qualty, Davd Chapman, Copper Development Assocaton, July Add to ths new demands on our dstrbuton network based on emergng technologes, and the added costs of outages for energy consumers can have even greater economc mpact. B. Dstrbuton Network Operators Because of the trend n ncreasng consumer costs due to supply nterrupton, dstrbuton network operators wll be forced to ncreasngly nvest more to mantan the optmal supply qualty level. One such soluton to ncrease supply qualty s Fault Locaton, Isolaton and Servce Restoraton (FLISR). Ths soluton can easly help reduce the SAIDI metrc by allowng operators to locate and react to outages much more quckly. It can also help to lessen OPEX expendtures by reducng the amount of tme crews are requred to try and manually locate faults on the dstrbuton network. Further, functonalty to automate the servce restoraton process can restore many customers before the 1-3 mnute lmt, enablng reducton n both SAIDI and SAIFI. The remander of ths paper wll focus on evolutonary FLISR strateges that can be mplemented by dstrbuton network operators wth ncreasng CAPEX nvestment, as needed, to meet ncreasng supply qualty requrements. IV. FAULT PASSAGE INDICATORS The tradtonal method for locatng faults on the dstrbuton network can be a very tedous task. Crews are dspatched to the faulted feeder and try to predct the fault locaton. Usng a ht and mss method of solatng sectons of the faulted feeder, crews wll reclose the substaton breaker, or other network protectve devces, such as reclosers, to valdate that the faulted feeder secton s downstream from the solaton pont. Ths can create unnecessary wear on protectve equpment on the feeder beng tested, resultng n decreased equpment lfe and ncreased tme requred for mantenance. It also ncreases the rsk of creatng addtonal faults on the feeder as weak solaton ponts are stressed. Multple faults on a feeder can create

3 3 addtonal complexty for the crew as they try to locate the faulted secton. One soluton to help crews locate a fault and rule out healthy sectons of a feeder s the addton of fault passage ndcators (FPI) at strategcally located ponts along a feeder. Solutons for fault passage ndcaton exst for both overhead and underground networks. One such soluton for overhead networks s FPIs that can clp-on to overhead lnes. The FPIs operate by montorng the current at a specfc pont on the lne. To detect a fault, they montor the lne for a loss of voltage followng a sudden hgh mpulse current (di/dt) over a 30ms perod, or an overcurrent condton set to a preset value. Upon detectng a fault downstream from the FPI, the FPI wll provde a vsble ndcaton by flashng a set of hgh-ntensty LEDs. The crews can then follow the feeder and the flashng FPIs untl they locate an nactvated FPI (one that has not actvated the flashng LEDs). The crew wll then know that the fault s located between the last actve and frst nactve FPI, allowng them to focus ther efforts on a much smaller feeder secton. Fg 2. Overhead clp-on fault passage ndcators. SAIDI values can be sgnfcantly reduced by helpng crews locate faults and reducng the tme requred for dagnostc testng. Mantenance costs for protectve devces n the network can also be reduced by elmnatng the need to repeatedly close onto faults. FPIs can easly be added to the dstrbuton network wth very mnmal CAPEX nvestment, makng them an excellent startng pont for helpng operators to mprove ther qualty of supply. The ROI for addng FPIs can usually be realzed n 6-12 months. V. ADDING COMMUNICATIONS Typcally, communcatons has only extended to the dstrbuton substatons, allowng operators the ablty to detect faults that trp feeder breakers only. Ths leaves consumers to report outages caused by faults that operate other protectve devces outsde the substaton fence, such as reclosers or fuses. Ths means that the tme t takes to detect a fault s relant upon the consumer. Decdng on the rght communcatons medum depends on many unque factors, such as physcal or geographcal lmtatons that are specfc to dfferent dstrbuton network operators. Some of the many choces for communcaton medums nclude lcensed/unlcensed rado, cellular technology, broadband over power lne, fber optc cablng or even pggy-backng on mesh networks that are beng mplemented for Advanced Meterng Infrastructure (AMI). Addng communcatons capabltes to dstrbuton network devces outsde of the substaton fence can lead to mprovements n real-tme stuatonal awareness and even allow for some remote control capabltes. Devces exst today that can easly communcate to a network operator s supervsory control and data acquston (SCADA) system usng standard protocols such as DNP3 or MODBUS. Some of the benefts are specfcally related to communcatng fault passage ndcators and protectve/solaton devces such as reclosers and sectonalzers. A. Communcatng Fault Passage Indcators Communcatng FPIs operate along the same prncples as non-communcatng FPIs, except that upon the loss of voltage on a feeder; the outage can be mmedately reported to the dstrbuton network operator. Because the communcatng FPIs can also transmt nformaton f they have detected the passage of a fault current, the operator can determne the general locaton of a fault. The operator can quckly dspatch a crew drectly to the faulted secton of the feeder, sgnfcantly reducng the tme t takes to report and locate the fault on the feeder, whle also havng a mnmzng effect on the operator s SAIDI value. The ROI for communcatng FPIs can be realzed n 1-2 years. As an added beneft, communcatng FPIs can also provde meterng nformaton on average, mnmum and maxmum current values for the lne. Ths nformaton could be used to help optmze an operator s asset management program or to help detect dstrbuton network rregulartes.

4 4 Fg 3. Overhead communcatng fault passage ndcators. One such overhead communcatng FPI soluton today employs a pole mounted remote termnal unt (RTU) that communcates wth three or more FPIs through rado communcatons. There are some addtonal soluton health benefts from usng an RTU as a recever, such as the ablty to montor FPI sgnal strength or battery level. Ths nformaton can be reported back to an operator who can schedule mantenance at the frst sgn of potental trouble. The RTU can also be used to concentrate the FPI sensors to help optmze communcaton costs by reducng the number of cellular subscrptons or rado unts requred. B. Reclosers and Sectonalzers SCADA communcatons can also be extended to many exstng reclosers and sectonalzers n the dstrbuton network by nterfacng communcatons technology such as rados to the exstng controller. Ths wll not only allow mproved stuatonal awareness by reportng swtchgear statuses, currents and voltages, but also allow for remote control operaton of the swtchgear. In some cases, legacy recloser controls may not have the capablty to communcate through conventonal means or, n the case of some sectonalzers, not be equpped wth a controller. In these cases an ntegrated RTU (RTU) can be retroftted to the swtchgear ether through an output contact to an exstng controller or drectly. Some RTUs contan backup power supples that can provde enough power to operate sectonalzers and allow for control of the swtchgear durng an outage. Other RTU features nclude the ablty to nclude current transformer (CT) and voltage transformer (VT) nputs for meterng, and ntegrated communcatons technology such as rados or cellular communcatons. Lke the montorng of the feeder breakers at the substaton, addng the communcatons nfrastructure outsde the substaton allows dstrbuton network operators to mmedately be nformed when one of the protectve devces, such as reclosers, has operated. Data reported by the swtchgear controllers and RTUs on the feeder, along wth FPIs can ad n quckly locatng fault occurrences. Wth the operator s ablty to control the feeder swtchgear, the operator can manually solate the faulted area of the feeder by sendng control commands to reclosers and sectonalzers. Once the faulted area on the feeder s solated, the operator can remotely restore servce to the remander of the feeder consumers by closng the breaker or recloser upstream of the solated segment and by closng any possble normally open ponts downstream of the solated segment. Whle communcatng wth reclosers and sectonalzers does ad n fault locaton, the added value can be realzed n remotely solatng the fault and restorng servce to some of the consumers on an affected feeder, whle quckly mnmzng the mpact on consumers and the dstrbuton network operator s SAIDI metrc. VI. ADDING REGIONAL DISTRIBUTED INTELLIGENCE The next step after nstallng communcatons capabltes to the many ntellgent devces on the dstrbuton network s to add some dstrbuted ntellgence at a regonal level. One such method for accomplshng ths s to place an ntellgent controller at the dstrbuton substaton level. Ths regonal controller can contan a real-tme data model wth data collected from the ntellgent devces on the feeders connected to the substaton, dstrbuted generaton and neghborng substatons. The ntellgent controller would be relant on communcatons to regonal devces to adequately manage the regon, whle the controller should be able to act ndependently and not have relance on any centralzed ntellgence from the dstrbuton network level. Because of the regonal awareness the ntellgent controller possesses, t can easly assst n better determnng fault locaton, makng some automatc swtchng decsons and managng dstrbuted generaton connectons. A. Fault Locaton Determnaton Many rural feeders are radal and cover long dstances. Ths means the dstance between the placement of FPIs on the dstrbuton feeder can be large, requrng crews to spend greater amounts of tme searchng the feeder to locate the fault. Snce regonal ntellgent controllers have awareness of real-tme data such as voltages and currents collected from protectve devces nsde and outsde the substaton, they can help to more accurately predct the locaton of the fault usng the fault current method. The fault current method for detectng the locaton of faults uses the magntude of

5 5 the short-crcut current durng the last cycle before the protectve devce operates and the type of fault detected. From ths nformaton, a fault smulaton can be performed usng the pre-programmed parameters for the feeders to predct the locaton of the fault. In urban areas, ths method of fault locaton predcton should be used n conjuncton wth fault passage ndcators due to the added complexty of the network and varous feeder branches whch could lead to the fault current method predctng multple fault locatons. B. Automatc Swtchng After determnng the locaton of the fault on the feeder, the ntellgent controller can take some actons to ssue controls to sectonalzers to solate the fault as tghtly as possble. Once solaton s completed, the controller can restore servce to consumers upstream of the solated feeder secton. In some cases the controller could even automatcally restore servce to consumers downstream of the solated secton by closng a normally open pont between two adjonng feeders. Care should be taken before allowng ths type of operaton to ensure there s no possblty of overloadng the adjonng feeder. C. Dstrbuted Generaton Management Wth the new regulated requrement of dstrbuted generators connectng to feeders, multple supply sources need to be dsconnected from the feeder when a fault occurs. Snce the ntellgent controller s aware of the status of all protectve elements along a gven feeder, t can ssue dsconnect commands to dstrbuted generators downstream from the operated devce as per the dstrbuton network operator s grd code. VII. ADDING CENTRALIZED INTELLIGENCE Centralzed ntellgence, such as a dstrbuton management system (DMS), can be mplemented wth a real-tme awareness of the complete dstrbuton network. All decsons are made based on the current state of the network nstead of usng a worst-case, preengneered soluton. Detaled models also provde load smulatons that can assst wth detectng faults n the dstrbuton network. The DMS s aware of pre-fault loadng plus outage duraton and can predct the cold load pckup current expected when re-energzng the feeder secton. Ths allows for enhanced solaton and servce restoraton algorthms whch can be executed as a man n the mddle approach wth the SCADA operator nteractng wth the suggested restoraton steps, or fully automatc wth the restoraton steps executed n a closed loop control method usng fully SCADAcontrolled equpment. The DMS can produce restoraton swtchng steps as a result of a fault or can be used to mplement and restore servce after planned outages. Fg 4. Predcted restoraton load and optons for resupply. A. Fault Locaton For protectve devces that are dffcult to montor, such as fuses, smulaton of the expected loadng on a feeder can help to predct when and where these devces have operated. Ths, along wth data from customer reportng calls and nformaton on planned outages, can be used to help operators hone ther ablty to locate faults. B. Fault Isolaton and Servce Restoraton Wth the DMS system havng a global overvew of the entre network, the system can easly determne optmal swtchng routnes that can restore the maxmum number of consumers whle takng nto consderaton prorty customers and the ratng of specfc feeders. Ths global vew s mportant to ensure the transfer of load from one feeder to another does not cause overloadng on the feeder that s pckng up the load. For feeders that contan solaton devces that don t have remote control capabltes, the DMS can provde a lst of swtchng operatons that can be performed manually by the dspatched crew. The DMS system can also compute a lst of alternatve swtchng scenaros for operator selecton. The FLISR algorthms of the DMS can be ntated manually by the operator to reconfgure feeders due to planned outages or scheduled mantenance work. When feeders are operatng n non-standard confguratons that are outsde of normal operatons, the

6 6 DMS system can block the dstrbuted automatc feeder reconfguraton at the regonal level and ssue swtchng orders from the DMS level only. The result s fully coordnated, central-decentralzed ntellgence. C. Future Consderatons Many of the ntellgent devces n place today n the dstrbuton network have the ablty to record wave form captures of detected dstrbuton network faults. Future DMS systems wll have the capablty to collect and analyze these wave form captures and predct the cause of dfferent types of faults based on specfc fault sgnatures. For example, based on the fault sgnature of a tree branch fallng on a feeder, the operator would mmedately know that he also needs to dspatch a forestry crew to a specfc fault locaton. Future predcton tools wll provde an analyss of trends of transent events to help predct future falure modes. Dstrbuton network operators can then use ths nformaton to provde warnng of future outages and schedule mantenance before an outage or fault occurs. [5] Fault Locaton, Isolaton and Supply Restoraton, Techncal Specfcaton, Telvent DMS, February [6] Telvent DMS Functonal Overvew 2.0, Techncal Specfcaton, Telvent DMS, January Adam F. Gauc was born n Toronto, Ontaro, Canada and receved a Bachelors of Scence n computer engneerng from Queen s Unversty at Kngston, Ontaro. Hs prevous work experence ncludes Hydro One Networks as a Protecton and Control Engneer and Cooper Power Systems as a Feld Applcaton Engneer. Currently he s workng wth Smart Grd Solutons at the Schneder Electrc North Amercan Energy Automaton Center n Toronto, Ontaro. Mr. Gauc s currently a member of the IEEE Power and Energy Socety and a regstered professonal engneer n the provnce of Ontaro. VIII. CONCLUSION Wth new technologes and consumer requrements placng greater demand on the dstrbuton grd, addtonal pressure from consumers and regulators s beng felt by dstrbuton network operators to mprove the level of power supply qualty. Provdng greater relablty can be accomplshed by mplementng dfferent Fault Locaton, Isolaton and Servce Restoraton strateges whch can be measured by a relatve reducton n the operator s SAIDI metrc. Ths paper hghlghts some of the FLISR solutons that can be used to help reduce the SAIDI metrc wth varyng degrees of CAPEX nvestment. Investments n FLISR can easly evolve over tme and are scalable n terms of cost and complexty as new requrements demand. Addtonal operatonal and mantenance effcences can also be recognzed from nvestments n FLISR strateges that can also help justfy the CAPEX nvestments. REFERENCES [1] Y. Chollot, J.M. Basse and A. Mallot, Feeder Automaton Improves Medum Voltage Network Effcency, presented at CIRED Conference, Frankfurt, Germany, June , [2] M.J. Domngue and J. Chaves, News n Fault Passage Indcators n Overhead and Underground MV Lnes, presented at CIRED Conference, Barcelona, Span, May 12-15, [3] J.D. Kueck, B.J. Krby, P.N. Overholt and L.C. Markel, Measurement Practces for Relablty and Power Qualty, Oak Rdge Natonal Laboratory, Oak Rdge, TN Rep. ORNL/TM-2004/91, [4] K. Keller and B.F.C. Franken, "Qualty of Supply Market Regulaton Survey wthn Europe," Kema Labratores, Arnhem, Netherlands Rep TDC A, 2006.