SAIDI MINIMIZATION OF A REMOTE DISTRIBUTION FEEDER. Kai Zou, W. W. L. Keerthipala and S. Perera

Size: px

Start display at page:

Download "SAIDI MINIMIZATION OF A REMOTE DISTRIBUTION FEEDER. Kai Zou, W. W. L. Keerthipala and S. Perera"

Clare Anissa Powers
5 years ago
Views:

1 SAIDI INIIZATIN F A RETE DISTRIBUTIN FEEDER Ka Zou, W. W.. Keerthpala ad S. Perera Uversty of Wollogog School of Electrcal ad Computer Telecommucato Egeerg Wollogog, NSW 2522, Australa Abstract Dstrbuto system relablty assessmet s a mportat part of dstrbuto system operato ad plag. Ths paper presets some approaches take to mmze the SAIDI (System Average Iterrupto Durato Idex) for a remote dstrbuto feeder a kv dstrbuto system. Resultg from uavalablty of alteratve backup supples from the other adjacet feeders, dstrbuted geerators (DG) ca be used as backup supples ths case. Coordatg dstrbuto automato (DA) system wth DG, the restorato area ca be made larger ad the outage tme ca be shorter, cosequetly, the dstrbuto system relablty ca be mproved sgfcatly. I ths paper, factors affectg SAIDI are dscussed ad methods cludg feeder recofgurato, recloser stallato ad replacemet ad DG stallato are appled to mmze SAIDI. Comparatve studes are performed ad related results are addressed.. Itroducto Dstrbuto system relablty s the ablty of the dstrbuto system to perform ts fucto uder stated codtos for a stated perod of tme wthout falure []. Relablty mprovemet of electrc dstrbuto systems has bee a object of research efforts over may years. Relablty dces for dstrbuto systems have bee defed by several groups to set bechmarks power system desg. The power system desg ad mateace programs performed by utltes should deped o the relablty dces order to mprove system relablty effectvely. I rural areas, lack of alteratve backup supples from other adjacet feeders s a obstacle to havg hghly relable power supply these areas. Buldg a ew tercoecto feeder remote area s qute tme-cosumg ad cost-effectve due to complex geographcal evromet. But wth the developmet of dstrbuto automato (DA) ad dstrbuted geerato (DG), the dstrbuto systems located rural areas ca have hgher relablty at relatvely lower operato ad mateace costs. DGs are flexble to stall ad cost-effcet to operate ad mata. DGs ca also be regarded as backup supples durg terruptos. Approprate coordato of DA ad DG has great mpact o terrupto durato ad frequecy, whch sgfcatly affect the system relablty. I ths paper, some approaches take to assess the relablty of a remote dstrbuto system quattatvely are dscussed. I Secto 2, a remote dstrbuto feeder uder study s troduced ad aalysed. Factors affectg SAIDI are dscussed ad approaches to mmze SAIDI cludg feeder recofgurato, recloser replacemet ad DG stallato appled are preseted Secto 3. Smulato ad related results are Secto 4 followed by coclusos Secto Feeder Uder Study The dstrbuto feeder to be aalysed s a actual kv feeder located NSW. The data of ths feeder gve ths paper s modfed. The topology of ths feeder s show Fgure. It s assumed that the legth of each feeder secto s 2 km. The feeder sectos from to 4 are the ma feeder protected by a crcut breaker at the begg of ths feeder. The feeder sectos from 5 to 7 are the brach sectos protected by a fuse coected at the begg of ther related brach. The feeder sectos from 8 to 3 are the exteded brach sectos represetg creasg umber of customers ths area. For smplcty, the compoets such as feeder sectos, maual swtches ad trasformers each secto are combed together as oe compoet by usg relablty-etwork-equvalet method [2] ad the load pots are also lumped together each secto. It s also assumed that customers are uformly dstrbuted ad there are 5 customers each secto of the feeder. Ths dstrbuto feeder s suppled by a kv busbar ad o backup supply s avalable for ay secto of ths feeder resultg from the geographcal locato of the etre feeder. 3. Factors Affectg SAIDI 342

2 Fgure 2 Topology of a Smple Feeder Fgure Topology of the Remote Dstrbuto Feeder The objectve of ths paper s to mmze the dstrbuto system relablty dex SAIDI for feeder uder study. SAIDI ca be calculated eq.() [3]: SAIDI T = T r N () N T = Total umber of feeder sectos r = utage tme for each terrupto at N = The umber of terrupted customers = oad pot To calculate the outage tme for each load pot relato to dfferet scearos durg outages, Repar ad Isolato Tme atrx (RIT) as gve by eq.(2) s used. RIT = ( m ( m m + ) + m mm mm m ) ) ) ( ) ( ) j = utage tme for j whe fault m = Total umber of ma feeder sectos = Total umber of brach sectos (2) The square RIT ca be dvded to 4 regos. The top-rght rego dcates the outage tme o the ma feeder secto for faults the other ma feeder secto j. The top-left rego shows the outage tme o healthy ma feeder for faults o braches. Smlarly, the bottom-left rego shows the outage tme for each brach secto whe there are faults o ma feeder sectos ad the bottom-rght rego dcates the outage tme for each brach secto whe there are faults o the other brach sectos. utage tme could be ether repar tme or solato tme depedg o dfferet fault locatos, topology ad cofgurato of the feeder ad protecto scheme [4]. From eq.() ad eq.(2), SAIDI ca be calculated as follows: SAIDI = ( λ ) N j j = N j = utage tme for j whe fault o λ = Fault rate for secto m = Total umber of ma feeder sectos = Total umber of brach sectos N = The umber of terrupted customers = oad pot (3) Form eq.(3), t s clear that three parameters (outage tme for each secto, fault rate for each secto ad the umber of customers) ca affect SAIDI drectly. Sce the total umber of customers caot be chaged ths feeder ad the heret characterstcs of ths feeder makes the fault rate costat over a log-term perod [5], therefore, ths paper oly dscusses the factors affectg outage tme. Factors affectg outage tme are dscussed the followg sub-sectos. 3. Feeder Cofgurato The physcal feeder topology caot be chaged easly due to hgh cost ad geographcal dstrbuto 343

3 Table RIT for the feeder show Fgure 2(a) Feeder Sectos S S2 S3 S4 S5 S S S S S Table 2 RIT for the feeder show Fgure 2(b) Feeder Sectos S S2 S3 S4 S5 S S S S S Table 3 RIT for the feeder show Fgure 3 Feeder Sectos S S2 S3 S4 S5 S S S S S of customers, but dfferet protecto schemes ca chage the cofgurato of a feeder [6]. For stace, the cofgurato of the feeder show Fgure 2 ca be chaged by adjustg the breaker settg ad removg the fuse stalled at the begg of feeder secto 5 to form a ew ma feeder ad a fuse ca be stalled at the begg of feeder secto 3 to form a ew brach. Though feeder cofgurato ca be recofgured easly, the feeder capacty lmts ad thermal lmts should be cosdered. It s assumed that the fault rate for each secto Fgure 2 s 0.05 faults per year ad there are 0 customers served by each secto. The repar tme ad solato tme s assumed to be 6 hours ad 2 hours respectvely. The RITs for Fgure 2(a) ad Fgure 2(b) s show Table ad Table 2 respectvely. I both tables, the colums are the fault locatos ad the rows are the related outage tme for each secto. The SAIDI for the feeder Fgure 2(a) s 0.88 hour per year ad SAIDI for the feeder Fgure 2(b) s 0.90 hour per year, showg a slght reducto relablty levels. However, the recofgured feeder wll exhbt mproved relablty ad hece the SAIDI ca be reduced. 3.2 Protecto Scheme As metoed above, dfferet protecto schemes ca chage the feeder cofgurato. Not oly that, dfferet protecto schemes ca also affect the outage tme for dfferet feeder sectos. Fgure 3 Topology of a Smple Feeder wth AR Table 4 RIT for the feeder wth DG Fgure 3 Feeder Sectos S S2 S3 S4 S5 S S S S S A dstrbuto feeder wth suffcet DA should be more relable ad the average outage tme for a specfc perod should be shorter. But the cost of mplemetg DA has a trade-off relatoshp wth the relablty ssues. To have relatvely hgher relablty performace whle reducg the captal vestmet s very mportat for utltes [7, 8]. A example s llustrated to expla the effect of the protecto devces. For the same feeder show Fgure 2(a), t s supposed that a ew addtoal automatc recloser s to be stalled at the begg of feeder secto 3 as show Fgure 3. I ths case, the ew RIT s gve Table 3 ad the calculated SAIDI s 0.82 hour per year. The system relablty s mproved due to addtoal automatc recloser. The outage tme for every secto wll be more dffcult to determe f large umber of automatc reclosers s appled. I Secto 4, the effect of dfferet locatos ad umber of automatc reclosers o system relablty wll be compared detal. 3.3 Avalablty of Backup Supply I urba areas, backup supples are ot cosdered as a problem due to hghly meshed etworks. But remote areas, where oe dstrbuto feeder s far away from the other dstrbuto feeders, the backup supples from other feeders become uavalable. To buld a tercoecto feeder as a backup supply of the exstg feeder s expesve ad cost-effcet. I ths case, order to mprove system relablty, use of DGs s a good soluto. Nowadays, DGs are wdely used load sheddg ad voltage support dstrbuto systems. DGs are cosdered to be 344

4 Case Table 5 Descrpto of Case Studes Feeder Recofgured ptmze Recloser ocato Number of Reclosers DG Istallato DG ocato S S S S CASE- CASE-2 CASE-3 CASE-4 CASE-5 CASE-6 Fgure 4 Results of Case Studes CASE- CASE-2 CASE-3 CASE-4 CASE-5 CASE-6 relable backup supples whe part of the dstrbuto system exhbts a outage [7, 8]. Coordatg wth DA, the restorato area ca be larger ad the restorato process ca be much faster [0]. Table 4 shows the RIT whe a DG s stalled at the ed of feeder secto 4 show Fgure 3. It s assumed that the capacty of DG s suffcet to supply the whole feeder. Compared wth the SAIDI of feeders show Fgure 2 ad Fgure 3, the value of SAIDI ths case ca go further dow to 0.5 hour per year, whch proves that approprate stallato of DGs ca result better system relablty. ore cases about the mplemetato of DG for the feeder show Fgure wll be dscussed Secto 4 ad the assocated SAIDI values are also calculated Secto Case Studes 4. Assumptos To evaluate the relablty beeft mplemetg dfferet protecto schemes ad employg DGs as backup supples, dfferet comparatve studes are coducted. The assumptos behd varous case studes are descrbed as followg: Protecto devces: Protecto devces cludg automatc recloser, fuse ad crcut breaker are used each of the case studes. All protecto devces are assumed to be 00% relable all the tme, whch meas that all the protecto devces ca trp faults whe they are eeded. So the fault rate for protecto devces s 0 faults per year. Repar tme: Repar tme s the tme take to repar or replace the faulted electrc compoets. Repar tme ca vary depedg o several factors such as dfferet faulted compoets, weather codto ad fault locato. For smplcty, Repar tme case studes s assumed to be 6 hours for each faulted secto regardless of dfferet codtos. Isolato tme: Isolato tme ths paper s defed as the tme take to locate the fault secto ad the tme take to solate the fault secto from system. Isolato tme also ca vary depedg o the factors such as travel tme, avalable umber of crews, fault locato ad dfferet types of compoets. I ths paper, the solato tme s 2 hours for each secto. Fault rate: Fault rate for each feeder secto s assumed to be faults per year per km. I ths paper, oly the fault rate of feeder s cosdered because oly the faults o kv feeder sectos ca affect other healthy sectos. Fuses: All fuses are assumed to be stalled at the begg of the braches ad they are 00% relable. It s also assumed that the fuses ca oly trp the faults located o the brach sde ad all the fuses have bee set ad coordated wth other protecto devces properly. Automatc recloser: There are two types of automatc reclosers used case studes. e s the udrectoal automatc recloser whch ca oly detect the faults located dowstream; aother s the bdrectoal automatc recloser whch ca detect the faults o both sdes. It s assumed that all automatc reclosers ca coordate correctly to solate the faults as eeded. Dstrbuted geerator: I case studes, t s assumed that the capacty of DG s suffcet to supply the whole feeder ad the DG ca ru sladg mode whe requred. I ths paper, all DGs are oly started to supply the feeder sectos whch are outage subject to a emergecy mode. Due to the feeder capacty lmt, the DG stalled brach sectos ca oly supply related brach, but DG stalled ma feeder sectos ca supply both ma feeder ad brach sectos. 4.2 Comparatve Case Studes Descrptos of these comparatve case studes are show Table 5. A overall bref descrpto of 6 case studes s as follows: Case : Number of automatc reclosers s placed alog the feeder ad the recloser 345

5 placemet has bee optmzed to mmze SAIDI. The feeder cofgurato s ot chaged ad DGs are ot appled ths case. Case 2 ad 3: Number of automatc reclosers ad ther placemet are cosdered. DGs are troduced the two cases. e DG s placed at the ed of ma feeder Case 2 ad oe DG s placed at the ed of the brach Case 3. Case 4: The feeder s recofgured ths case. The fuse at the begg of feeder secto 5 s removed ad a ew fuse s placed at the begg of feeder secto 6 to form a ew brach. Number of automatc reclosers s cosdered ad ther placemet s optmzed. Case 5 ad 6: DGs are appled both two cases for the recofgured feeder. The dfferece s that the dstrbuted geerator s stalled at ed of ma feeder Case 5 whle Case 6 the dstrbuted geerator s stalled at the ed of brach. 4.3 Results of Case Studes The results of 6 comparatve case studes are llustrated Fgure 4. It s otced that f the feeder s recofgured wth automatc reclosers or DGs, the system relablty wll become worse for ths feeder. The value of SAIDI wll crease slghtly, from 737 mutes Case to 749 mutes Case 4, f oly feeder recofgurato s appled. Comparg Case ad Case 2 wth Case 4 ad Case 5, t s obvous that the values of SAIDI are reduced f automatc reclosers are placed properly ad the DG s stalled at the ed of ma feeder whe the feeder s recofgured. But Case 3 ad Case 6, whe the dstrbuted geerator s placed at the ed of the brach, the value of SAIDI wll crease f the feeder s recofgured. It s also observed from Fgure 4 that the value of SAIDI decreases wth the crease of the umber of automatc reclosers. But the rate of decrease of SAIDI becomes smaller whe the umber of automatc reclosers s creased. Cosequetly, t s suggested that the utltes should cosder a tradeoff betwee captal cost ad relablty mprovemet. I comparatve case studes, t s dcated that Case 3 gves the best results. Eve whe o automatc recloser s appled, the maxmum value of SAIDI Case 3 ca be half of the value Case. If 5 automatc reclosers are stalled, the value of SAIDI ca reach ts lowest value of 64 mutes. Smlar values are obtaed for Case Coclusos Ths paper preseted some approaches for the SAIDI mmzato of a remote dstrbuto feeder a kv dstrbuto system. The approaches are based o the relablty-etwork-equvalet model whch the mpacts of feeder cofgurato, automatc recloser placemet ad dstrbuted geerator applcato are cluded the aalyses. The factors affectg system relablty were also dcated ths paper. Comparatve case studes are performed to evaluate the system relablty mprovemet by usg dfferet methods. The study results llustrated ths paper dcate that the system relablty s mproved by applyg a dstrbuted geerator as a backup supply. Coordatg automatc recloser wth dstrbuted geerator, the system relablty ca be further mproved. I ths paper, DG was assumed to be completely defed. However, future research s requred optmzg the locato ad sttg of dfferet types of dstrbuted geerators. 6. Refereces [] Cheryl A. Warre, Dstrbuto Relablty - What s t? IEEE Idustry Applcatos agaze, Vol. 2, Issue 4, July-Aug 996, pp [2] R.Bllto ad P.Wag, Relablty-etworkequvalet approach to dstrbuto-systemrelablty evaluato'', IEEE Proceedgs. Geerato. Trasmsso ad Dstrbuto, Vol. 45, No. 2, arch 998, pp [3] IEEE gude for electrc power dstrbuto relablty dces, IEEE , [4] Carr.W. Predctve dstrbuto relablty aalyss cosderg post fault restorato ad coordato falure, Rural Electrc Power Coferece, IEEE, 5-7 ay 2002, Pages: B3-B3_6. [5] R.Bllto ad R.N.Alla, Relablty Evaluato of Power Systems, 2d ed. Pleum, New York, 996. [6] R.E.Brow, Electrc Power Dstrbuto Relablty. arcel Dekker, New York, 2002 [7] A.S.Pabla, Electrc Power Dstrbuto. cgraw-hll, New York, 2004 [8] cdermott, T.E.Duga, R.C, PQ, relablty ad DG, Idustry Applcato agaze, IEEE, Vol.9, Iss.5, pp.7-23, Sept.-ct [9].H.J.Bolle, Y.Su, G.W.Ault, Relablty of Dstrbuto Networks wth DER cludg Itetoal Isladg, Future Power Systems, 2005 Iteratoal Coferece o, pp.-6, 6-8 Nov [0] S.Kazem,.Fotuh-Fruzabad, R.Bllto, Relablty assessmet of a automated dstrbuto system, Geerato, Trasmsso & Dstrbuto, IET, Vol., pp , arch

Saidi minimization of a remote distribution feeder

Saidi minimization of a remote distribution feeder Unversty of Wollongong Research nlne Faculty of Informatcs - Papers (Archve) Faculty of Engneerng and Informaton Scences 2007 Sad mnmzaton of a remote dstrbuton feeder Ka Zou Unversty of Wollongong, kz965@uow.edu.au