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 Wckramaarachch W. Keerthpala Unversty of Wollongong, wkeerth@uow.edu.au Sarath Perera Unversty of Wollongong, sarath@uow.edu.au Publcaton Detals K. Zou, W. W. Keerthpala & S. Perera, "Sad mnmzaton of a remote dstrbuton feeder," n Conference Proceedngs of the 2007 Australasan Unverstes Power Engneerng Conference: AUPEC 2007, 2007, pp. 342-346. Research nlne s the open access nsttutonal repostory for the Unversty of Wollongong. For further nformaton contact the UW brary: research-pubs@uow.edu.au
Sad mnmzaton of a remote dstrbuton feeder Abstract Dstrbuton system relablty assessment s an mportant part of dstrbuton system operaton and plannng. Ths paper presents some approaches taken to mnmze the SAIDI (System Average Interrupton Duraton Index) for a remote dstrbuton feeder n an 11kV dstrbuton system. Resultng from unavalablty of alternatve backup supples from the other adjacent feeders, dstrbuted generators (DG) can be used as backup supples n ths case. Coordnatng dstrbuton automaton (DA) system wth DGs, the restoraton area can be made larger and the outage tme can be shorter, consequently, the dstrbuton system relablty can be mproved sgnfcantly. In ths paper, factors affectng SAIDI are dscussed and methods ncludng feeder reconfguraton, recloser nstallaton and replacement and DG nstallaton are appled to mnmze SAIDI. Comparatve studes are performed and related results are addressed. Keywords Sad, mnmzaton, remote, dstrbuton, feeder Dscplnes Physcal Scences and athematcs Publcaton Detals K. Zou, W. W. Keerthpala & S. Perera, "Sad mnmzaton of a remote dstrbuton feeder," n Conference Proceedngs of the 2007 Australasan Unverstes Power Engneerng Conference: AUPEC 2007, 2007, pp. 342-346. Ths conference paper s avalable at Research nlne: http://ro.uow.edu.au/nfopapers/1687
1 SAIDI nmzaton of a Remote Dstrbuton Feeder Ka Zou, W. W.. Keerthpala, and S. Perera Abstract Dstrbuton system relablty assessment s an mportant part of dstrbuton system operaton and plannng. Ths paper presents some approaches taken to mnmze the SAIDI (System Average Interrupton Duraton Index) for a remote dstrbuton feeder n an 11kV dstrbuton system. Resultng from unavalablty of alternatve backup supples from the other adjacent feeders, dstrbuted generators (DG) can be used as backup supples n ths case. Coordnatng dstrbuton automaton (DA) system wth DGs, the restoraton area can be made larger and the outage tme can be shorter, consequently, the dstrbuton system relablty can be mproved sgnfcantly. In ths paper, factors affectng SAIDI are dscussed and methods ncludng feeder reconfguraton, recloser nstallaton and replacement and DG nstallaton are appled to mnmze SAIDI. Comparatve studes are performed and related results are addressed. Index Terms Power dstrbuton relablty, power dstrbuton protecton, dstrbuted generator. D I. INTRDUCTIN ISTRIBUTIN relablty s the ablty of the dstrbuton system to perform ts functon under stated condtons for a stated perod of tme wthout falure [1]. Relablty mprovement of electrc dstrbuton systems has been an object of research efforts over many years. Relablty ndces for dstrbuton systems have been defned by several groups to set benchmarks n power system desgn. The power system desgn and mantenance programs performed by utltes should depend on the relablty ndces n order to mprove system relablty effectvely. In rural areas, lack of alternatve backup supples from other adjacent feeders s an obstacle to havng hghly relable supply n these areas. Buldng a new nterconnecton feeder n remote area s qute tme-consumng and cost-neffectve due to complex geographcal envronment. But wth the development of dstrbuton automaton (DA) and anuscrpt receved August 15, 2007; revsed ctober 20, 2007. Ka Zou s wth the School of Electrcal, Computer and Telecommuncatons Engneerng, Unversty of Wollongong, Wollongong, NSW, 2522 Australa (e-mal: kz965@uow.edu.au). W.W..Keerthpala s wth the School of Electrcal, Computer and Telecommuncatons Engneerng, Unversty of Wollongong, Wollongong, NSW, 2522 Australa (e-mal: wkeerth@uow.edu.au). S.Perera s wth the School of Electrcal, Computer and Telecommuncatons Engneerng, Unversty of Wollongong, Wollongong, NSW, 2522 Australa (e-mal: sarath@uow.edu.au). dstrbuted generaton (DG), the dstrbuton systems located n rural area can have hgher relablty at relatvely lower operaton and mantenance costs. DGs are flexble to nstall and cost-effcent to operate and mantenan. DGs can also be regarded as backup supples durng nterruptons. Approprate coordnaton of DA and DG has great mpact on nterrupton duraton and frequency, whch s sgnfcantly affect the system relablty. In ths paper, some approaches taken to assess the relablty of a remote dstrbuton system quanttatvely are dscussed. In Secton II, a remote dstrbuton feeder under study s ntroduced and analysed. Factors affectng SAIDI are dscussed and approaches to mnmze SAIDI ncludng feeder reconfguraton, recloser replacement and DG nstallaton are presented n Secton III. Smulaton and related results are n Secton IV followed by conclusons n Secton V. II. FEEDER UNER STUDY The dstrbuton feeder to be analysed s an 11kV feeder located n a remote area. The topology of ths feeder s shown n Fg. 1. It s assumed that the length of each secton s 2km. The feeder sectons from 1 to 14 are the man feeder protected by a breaker at the begnnng of ths feeder. The feeder sectons from 15 to 17 are the branch sectons protected by a fuse connected at the begnnng of ther related branch. The feeder sectons from 18 to 31 are the extended branch sectons representng ncreasng number of customers n ths area. For smplcty, the components such as feeder sectons, manual swtches and transformers n each secton are combned together as one component by usng relablty-networkequvalent method [2] and the load ponts are lumped together n each secton. It s also assumed that customers are unformly dstrbuted and there are 15 customers n each secton. Ths dstrbuton feeder s suppled by an 11kV busbar and no backup supply s avalable for any secton of ths feeder resultng from the geographcal locaton of the entre feeder. III. FACTRS AFFECTING SAIDI The objectve of ths paper s to mnmze the dstrbuton system relablty ndex SAIDI for feeder under study. SAIDI can be calculated by usng ( [3]:
2 Fg. 2. Topology of a Smple Feeder TABE I RIT FR THE FEEDER SHWN IN FIG. 2(A) Fg. 1. Topology of the Remote Dstrbuton Feeder SAIDI T = 1 = T = 1 r N N ( confguraton of the feeder and protecton scheme [4]. From ( and (2), SAIDI can be calculated as follows: where T s the total number of sectons, r s the outage tme for each nterrupton at, N s the number of nterrupted customers for each sustaned nterrupton durng the reported perod. To calculate the outage tme for each load pont n relaton to dfferent scenaros durng outages, Repar and Isolaton Tme atrx (RIT) s used n ths paper. RIT = ( m ( m 11 m1 + 1 + 1 1m mm mm ( m 1( ( ( 1( ( ( n where j s the outage tme for secton j when there s a fault on secton, m s the total number of man feeder sectons and n s the total number of branch sectons. The square RIT can be dvded nto 4 regons. The toprght regon (from 11 to mm) ndcates the outage tme on man feeder secton for faults on the other man feeder secton j. The top-left (from 1( to ) regon shows the outage tme on healthy man feeder for faults on branches. Smlarly, the bottom-left regon (from (1 to (m) shows the outage tme for each branch secton when there are faults on man feeder sectons and the bottom-rght regon (from (( to (() ndcates the outage tme for each branch secton when there are faults on the other branch sectons. utage tme could be ether repar tme or solaton tme dependng on dfferent fault locatons, topology and ) (2) SAIDI = n n ( λ ) N j = 1 j = 1 n = 1 N where λ s the fault rate for secton. Form (3), t s clear that three parameters (outage tme for each secton, fault rate for each secton and the number of customers) can affect SAIDI drectly. Snce the total number of customers cannot be changed n ths feeder and the nherent characterstcs of ths feeder makes the fault rate constant over a long-term perod [5], therefore, ths paper only dscusses the factors affectng outage tme. Factors affectng outage tme are dscussed n the followng sub-sectons: A. Feeder Confguraton The physcal feeder topology cannot be changed easly due to hgh cost and geographcal dstrbuton of customers, but dfferent protecton schemes can change the confguraton of a feeder [6]. For nstance, the confguraton of the feeder shown n Fg. 2 can be changed by adjustng the breaker settng and removng the fuse nstalled at the begnnng of feeder secton 5 to form a new man feeder and a fuse can be nstalled at the begnnng of feeder secton 3 to form a new branch. Though feeder confguraton can be reconfgured easly, however, the feeder capacty lmts and thermal lmts should be consdered. It s assumed that the fault rate for each secton n Fg. 2 s 0.05 faults per year and there are 10 customers served by each secton. The repar tme and solaton tme s assumed to be 6 (3)
3 TABE II RIT FR THE FEEDER SHWN IN FIG.2(B) TABE III RIT FR THE FEEDER SHWN IN FIG.3 TABE IV RIT FR THE FEEDER SHWN IN FIG. 3 Fg. 3. Topology of a Smple Feeder wth AR hours and 2 hours respectvely. The RITs for Fg. 2(a) and Fg. 2(b) s shown n Table I and Table II respectvely. In both tables, the columns are the fault locatons and the rows are the related outage tme for each secton. The SAIDI for the feeder n Fg. 2(a) s 0.88 hour per year and SAIDI for the feeder n Fg. 2(b) s 0.90 hour per year showng a slght reducton n relablty levels. However, the reconfgured feeder wll exhbt mproved relablty and hence the SAIDI can be reduced. B. Protecton Scheme As mentoned above, dfferent protecton schemes can change the feeder confguraton. Not only that, dfferent protecton schemes can also affect the outage tme for dfferent feeder sectons. A dstrbuton feeder wth suffcent DA should be more relable and the average outage tme for a specfc perod should be shorter. But the cost of mplementng DA has a trade-off relatonshp wth the relablty ssues. To have relatvely hgher relablty performance whle reducng the captal nvestment s very mportant for utltes [7, 8]. An example s llustrated to explan the effect of the protecton devces. For the same feeder shown n Fg. 2(a), t s supposed that a new automatc recloser s to be nstalled at the begnnng of feeder secton 3 as shown n Fg. 3. In ths case, the new RIT s gven n Table III and the calculated SAIDI s 0.82 hour per year. The system relablty s mproved due to addtonal automatc recloser. The outage tme wll be more dffcult to determne f large number of automatc reclosers s appled. In Secton IV, the effect of dfferent locatons and number of automatc reclosers on system relablty wll be compared n detal. C. Avalablty of Backup Supply In urban area, backup supples are not consdered as a problem due to hghly meshed network. But n remote area, where the dstrbuton feeder s far away from the other dstrbuton feeders, the backup supples from the other feeders become unavalable. To buld an nterconnecton feeder as a backup supply of the exstng feeder s expensve and costneffcent. In ths case, n order to mprove system relablty, use of DGs s a good soluton. Nowadays, DGs are wdely used n load sheddng and voltage support n dstrbuton systems. DGs are consdered to be relable backup supples when part of the dstrbuton system exhbts an outage [7, 8]. Coordnatng wth DA, the restoraton area can be larger and the restoraton process can be much faster [10]. Table IV shows the RIT when a DG s nstalled at the end of feeder secton 4 shown n Fg. 3. It s assumed that the capacty of DG s suffcent to supply the whole feeder. Compared wth the SAIDI of feeders shown n Fg. 2 and Fg. 3, the value of SAIDI n ths case can go further down to 0.5 hour per year, whch proves that approprate nstallaton of DGs can result n better system relablty. ore cases about the mplementaton of DG for the feeder shown n Fg. 1 wll be dscussed n Secton IV and the assocated SAIDI values are also calculated n Secton IV. A. Assumptons IV. CASE STUDIES To evaluate the relablty beneft of mplementng dfferent protecton schemes and employng DGs as backup supples, dfferent comparatve studes are conducted. The assumptons behnd n varous case studes are descrbed as followng:
4 Protecton Devces: Protecton devces ncludng automatc recloser, fuse and crcut breaker are used n case studes. All the protecton devces are assumed to be 100% relable all the tme, whch means that all the protecton devces can trp faults as they are needed. So the fault rate for protecton devces s 0 faults per year. 2) Repar Tme: Repar tme s the tme taken to repar or replace the faulted electrc components. Repar tme can vary dependng on several factors such as dfferent faulted electrc component, weather condton and fault locaton. For smplcty, repar tme n case studes s assumed to be 6 hours for each faulted secton regardless of dfferent condtons. 3) Isolaton Tme: Isolaton tme n ths paper s defned as the tme taken to locate the fault and the tme taken to solate the fault from system. Isolaton tme also can vary dependng on the factors such as travel tme, avalable number of crews, fault locaton and dfferent types of electrc components. In ths paper, the solaton tme s 2 hours for each secton. 4) Fault Rate: Fault rate for each feeder secton s assumed to be 0.066 faults per year per km. In ths paper, only the fault rate of feeder s consdered because only the faults on 11kV feeder sectons can affect other healthy sectons. 5) Fuse: All fuses are assumed to be nstalled at the begnnng of the branches and they are 100% relable. It s also assumed that the fuses can only trp the faults located on the branch sde and all the fuses have been set and coordnated wth other protecton devces properly. 6) Automatc Recloser: There are two types of automatc reclosers used n case studes. ne s the undrectonal automatc recloser whch can only detect the faults located downstream; another s the bdrectonal automatc recloser whch can detect the faults on both sdes. It s assumed that all automatc reclosers can coordnate correctly to solate the faults as needed. 7) Dstrbuted Generator: In case studes, t s assumed that the capacty of dstrbuted generator s suffcent to supply the whole feeder and the dstrbuted generator can run n slandng mode when requred. In ths paper, all DGs are only started to supply the feeder sectons whch are n outage subject to an emergency mode. Due to feeder capacty lmt, the DG nstalled n branch sectons can only supply related branch, but dstrbuted generator nstalled n man feeder sectons can supply both man feeder and branch sectons. B. Comparatve Case Studes Descrptons of these comparatve case studes are shown n Table V. An overall bref descrpton of 6 case studes s as follows: Case 1: Number of automatc reclosers s placed along the feeder and the recloser placement has been optmzed to mnmze SAIDI. The feeder confguraton s not changed and DGs are not appled n ths case. 2) Case 2 and 3: Number of automatc reclosers and ther placement are consdered. DGs are ntroduced n these two cases. ne DG s placed at the end of man feeder n Case 2 and one DG s placed s placed at the end of the branch n Fg. 4. Results of Case Studes TABE V DESCRIPTIN F CASE STUDIES Case 3. 3) Case 4: The feeder s reconfgured n ths case. The fuse at the begnnng of feeder secton 15 s removed and a new fuse s placed at the begnnng of feeder secton 6 to form a new branch. Number of automatc reclosers s consdered and ther placement s optmzed. 4) Case 5 and 6: DGs are appled n both two cases for the reconfgured feeder. The dfference s that the dstrbuted generator s nstalled at the end of man feeder n Case 5 whle n Case 6 the dstrbuted generator s nstalled at the end of branch. C. Results of Case Studes The results of 6 comparatve case studes are llustrated n Fg. 4. It s notced that f the feeder s reconfgured wth automatc reclosers or DGs, the system relablty wll become worse for ths feeder. The value of SAIDI wll ncrease slghtly, from 737 mnutes n Case 1 to 749 mnutes n Case 4, f only feeder reconfguraton s appled. Comparng Case 1 and Case 2 wth Case 4 and Case 5, t s obvous that the values of SAIDI are reduced f automatc reclosers are placed properly and the DG s nstalled at the end of man feeder when the feeder s reconfgured. But n Case 3 and Case 6, when the dstrbuted generator s placed at the end of the branch, the value of SAIDI wll ncrease f the feeder s reconfgured. It s also obvous from Fg. 4 that the value of SAIDI decreases wth the ncrease of the number of automatc reclosers. But the rate of decrease of SAIDI becomes smaller
5 when the number of automatc reclosers s ncreased. Consequently, t s suggested that the utltes should consder a tradeoff between captal cost and relablty mprovement. In comparatve case studes, t s ndcated that Case 3 gves the best results. Even when no automatc recloser s appled, the maxmum value of SAIDI n Case 3 can be half of the value n Case 1. If 5 automatc reclosers are nstalled, the value of SAIDI can reach ts lowest value of 164 mnutes. Smlar values are obtaned for Case 5. V. CNCUSINS Ths paper presented some approaches for the SAIDI mnmzaton of a remote dstrbuton feeder n an 11kV dstrbuton system. The approaches are based on the relablty-network-equvalent model n whch the mpacts of feeder confguraton, automatc recloser placement and dstrbuted generator applcaton are ncluded n the analyss. The factors affectng system relablty were also ndcated n ths paper. Comparatve case studes are performed to evaluate the system relablty mprovement by usng dfferent methods. The study results llustrated n ths paper ndcate that the system relablty s mproved by applyng a dstrbuted generator as a backup supply. Coordnatng automatc recloser wth dstrbuted generator, the system relablty can be further mproved. In ths paper, DG was assumed to be completely defned. However, future research s requred n optmzng the locaton and sttng of dfferent types of dstrbuted generators. REFERENCES [1] Cheryl A. Warren, Dstrbuton Relablty - What s t? IEEE Industry Applcatons agazne, Vol. 2, Issue 4, July-Aug 1996, pp. 32-37. [2] R.Bllnton and P.Wang, Relablty-network-equvalent approach to dstrbuton-system-relablty evaluaton'', IEEE Proceedngs. Generaton. Transmsson and Dstrbuton, Vol. 145, No. 2, arch 1998, pp. 149-153 [3] IEEE gude for electrc power dstrbuton relablty ndces, IEEE 1366-2003, 2004. [4] Carr.W. Predctve dstrbuton relablty analyss consderng post fault restoraton and coordnaton falure, Rural Electrc Power Conference, 2002. 2002 IEEE, 5-7 ay 2002, Pages: B3-B3_6. [5] R.Bllnton and R.N.Allan, Relablty Evaluaton of Power Systems, 2nd ed. Plenum, New York, 1996. [6] R.E.Brown, Electrc Power Dstrbuton Relablty. arcel Dekker, New York, 2002 [7] A.S.Pabla, Electrc Power Dstrbuton. cgraw-hll, New York, 2004 [8] cdermott, T.E.Dugan, R.C, PQ, relablty and DG, Industry Applcaton agazne, IEEE, Vol.9, Iss.5, pp.17-23, Sept.-ct. 2003. [9].H.J.Bollen, Y.Sun, G.W.Ault, Relablty of Dstrbuton Networks wth DER ncludng Intentonal Islandng, Future Power Systems, 2005 Internatonal Conference on, pp.1-6, 16-18 Nov. 2005. [10] S.Kazem,.Fotuh-Fruzabad, R.Bllnton, Relablty assessment of an automated dstrbuton system, Generaton, Transmsson & Dstrbuton, IET, Vol.1, pp.223-233, arch 2007.